Satellite perception assisted secure communication method based on dynamic uncertainty modeling
By using active detection and multi-dimensional channel modeling with ISAC dual-function beams, the channel uncertainty problem caused by unknown eavesdroppers in low-Earth orbit satellite communication is solved, realizing dynamic adaptive secure transmission and improving the security and communication efficiency of satellite communication systems.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHONGQING UNIV OF POSTS & TELECOMM
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-07
AI Technical Summary
When low-Earth orbit satellite communication systems face high-speed movement and unknown eavesdroppers, existing physical layer security technologies struggle to achieve dynamic and adaptive defense, making it difficult to obtain channel state information and effectively respond to security threats.
Active detection is performed using a dual-function beam based on ISAC. A multi-dimensional channel uncertainty model is established, and the optimal beamforming and receiving filter are jointly designed. The configuration is optimized through mathematical transformation and algorithm iteration to achieve dynamic adaptive secure transmission.
It enables proactive detection and dynamic defense against potential eavesdroppers, improves the security of satellite communication systems and the communication rate of legitimate users, and overcomes the shortcomings of fixed boundary settings in traditional methods.
Smart Images
Figure CN122349104A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of satellite-to-ground communication and physical layer security technology, specifically to a method for secure transmission of low Earth orbit (LEO) satellites based on dynamic uncertainty modeling, utilizing active radar sensing assistance, and combining worst-case optimization. Background Technology
[0002] With the continuous evolution of B5G / 6G communication technologies, integrated air-space-ground-sea wireless communication networks have attracted much attention due to their ability to provide seamless, high-speed, and reliable connectivity. Among them, satellite communication, with its inherent advantage of wide-area coverage, has become a core support for achieving global interconnectivity. Especially in areas where traditional terrestrial networks are difficult to reach, such as deserts, oceans, and remote rural areas, satellite communication plays an irreplaceable role. Compared with traditional geostationary orbit (GEO) satellites, low Earth orbit (LEO) satellites have significant advantages such as lower orbital altitude, lower transmission delay, lower path loss, and lower manufacturing costs, making them a research hotspot in academia and industry in recent years. However, the openness and wide coverage of satellite communication systems make downlink signals easily exposed in the vast physical space. Any ground node with receiving capabilities can intercept the radio signals broadcast by the satellite, posing a severe risk of information leakage and security threats. Therefore, how to ensure the secure transmission of sensitive data and private information in open satellite-to-ground links has become a key challenge that urgently needs to be solved in the large-scale commercialization of LEO satellite communication systems.
[0003] Traditionally, to ensure the security of satellite communications, systems typically rely on encryption technologies at the upper layers of the protocol stack. This approach assumes that an eavesdropper (Eve) has limited computing power and cannot crack complex encryption algorithms within a limited timeframe. However, with the rapid development of quantum computing and high-performance computing, eavesdroppers' signal processing capabilities are constantly increasing, posing unprecedented challenges to this computational security assumption based on computational complexity. To address this challenge, Physical Layer Security (PLS) technology has emerged and is gradually becoming another major direction for ensuring the security of satellite-to-ground links. PLS technology transcends the framework of traditional cryptography, directly utilizing the inherent physical characteristics of wireless channels (such as noise, fading, interference, and spatial degrees of freedom) to achieve secure information transmission. In low-Earth orbit satellite systems, the currently widely adopted technology is beamforming combined with multi-antenna arrays. By carefully designing the precoding matrix at the satellite transmitter, the system can highly concentrate and directionally transmit signal energy to legitimate ground users, while simultaneously creating nulls or reducing signal strength in the direction of potential eavesdroppers. This physically widens the capacity gap between legitimate and eavesdropping channels, achieving absolute security at the physical layer.
[0004] While beamforming-based physical layer security technology theoretically offers robust security, its practical application in low-Earth orbit (LEO) satellite communication scenarios faces extremely severe challenges. These challenges primarily stem from the extreme difficulty of obtaining accurate Channel State Information (CSI). First, for legitimate users, LEO satellites are in a state of high-speed movement, resulting in significant Doppler shift and rapidly changing channel characteristics. Limited by the long satellite-to-ground transmission delay, it is extremely difficult for the satellite to accurately obtain the real-time angle information of legitimate users, leading to beam pointing deviations. Second, obtaining the channel state information of eavesdroppers is even more challenging. In real-world wide-area open environments, eavesdroppers are typically passive, covert, and unverified illegal nodes that often remain silent when intercepting signals. This means that satellite communication systems cannot determine the exact physical location or perfect channel information of eavesdroppers in advance through traditional channel estimation or feedback mechanisms. This objective reality of "unknown eavesdroppers" makes the traditional perfect channel assumption extremely impractical in LEO satellite scenarios, severely limiting the effectiveness of secure beamforming technology.
[0005] To address the challenges posed by channel uncertainties, academia and industry have proposed a series of robust physical layer security methods under the general direction of beamforming. The first common approach is artificial noise (AN)-assisted techniques, which inject artificial interference noise into the null space of the legitimate channel while transmitting useful signals, aiming to degrade the quality of the eavesdropping channel. The second type is robust optimization methods based on statistical channel information, assuming that channel errors follow a known distribution, thereby optimizing the average secrecy rate. The third type is worst-case optimization methods for when the eavesdropper's location is completely unknown. The most typical approach is to use a norm-bounded model to define a potential error region for the eavesdropper and use this as a boundary condition for conservative beam design. However, these methods still have insurmountable bottlenecks and problems. First, regardless of the method used, due to the lack of proactive detection capability for eavesdropping targets, the system remains in a passive defense state against a "hypothetical enemy." Second, in robust designs using norm-bounded models, the uncertainty boundary is often a fixed, empirically based constant value. If the fixed boundary is set too large, the satellite will waste a significant amount of valuable transmission power on ineffective defenses, severely sacrificing the communication rate for legitimate users; if it is set too small, security vulnerabilities will arise when the actual error exceeds expectations. In other words, existing robust designs are rigid and lack environmental adaptability, failing to dynamically adjust according to the actual level of security threats. Summary of the Invention
[0006] To address the shortcomings of existing technologies, this invention proposes a sensing-assisted robust satellite secure transmission method based on integrated sensing, which includes:
[0007] S1: Construct an integrated signal transmission architecture;
[0008] S2: Actively detect and lock onto potential eavesdroppers;
[0009] S3: Establish a multidimensional channel uncertainty model;
[0010] S4: Jointly design the optimal beamforming and receiving filter;
[0011] S5: Mathematical transformations and iterative algorithm solutions.
[0012] Preferably, an integrated signal transmission architecture is constructed: the satellite end is based on an integrated communication and sensing architecture, using a unified hardware platform and a uniform planar antenna array to generate a dual-function beam that combines communication and sensing functions; during the downlink transmission phase, the dual-function beam transmits private data to legitimate users while also serving as a radar detection signal to cover potential threat areas.
[0013] Preferred methods for actively detecting and locating potential eavesdroppers include:
[0014] The radar receiver filter captures the echo signals reflected from ground targets.
[0015] By analyzing the azimuth, elevation, and time delay of the echo signal, the estimated center coordinates of the potential eavesdropper are calculated.
[0016] Based on the signal-to-noise ratio of the echo signal and the real-time motion state of the satellite, the positioning results are dynamically evaluated, and the uncertainty boundary of the positioning is calculated. .
[0017] Preferably, establishing a multidimensional channel uncertainty model includes:
[0018] To address the high-speed movement characteristics of low-Earth orbit satellites, a legitimate user channel angle uncertainty model incorporating angle deviation parameters is established to prevent beam drift.
[0019] Based on the estimated center coordinates and uncertainty boundary We construct a norm-bounded uncertainty region for the eavesdropper, which covers all possible physical locations where the eavesdropper may exist.
[0020] Preferably, the joint optimization problem of designing the optimal beamforming and constructing the receiving filter is expressed as:
[0021]
[0022]
[0023]
[0024] in, Indicates the first Worst-case confidentiality rate for each user This indicates the output signal-to-noise ratio of the radar detection beam. This represents the preset perception constraint threshold. Represented as the first A user-designed transmit beamforming precoding vector This represents the transmit power of the beamforming vector. This represents the budget ceiling for the total satellite launch power.
[0025] Furthermore, the first Worst-case secrecy rate for each user Defined as:
[0026]
[0027] in, This represents the array response estimation error vector of the eavesdropper. This represents the uncertainty boundary derived from dynamic evaluation. Indicates a legitimate user achievable data rates This indicates that the eavesdropper intercepted the first The achievable rate of a user signal, This indicates the operation of retrieving the maximum value. To ensure that the secrecy rate is non-negative.
[0028] Preferably, the process of obtaining the optimal configuration by using mathematical transformations and algorithmic iterations includes:
[0029] By introducing a positive semidefinite matrix, the complex beam vector non-convex optimization problem is transformed into matrix operations using positive semidefinite relaxation techniques.
[0030] A linearized approximation is performed on the non-convex subtractive term in the reachable data rate of legal users using a convex-concave process and a first-order Taylor expansion.
[0031] By combining the S-Procedure theorem, the non-convex security constraint containing the bounded error term of the eavesdropper norm is transformed into an equivalent finite-dimensional linear matrix inequality constraint;
[0032] Alternately solve the transformed linear matrix inequality convex optimization subproblem and the radar receiver filter optimization problem based on the generalized Rayleigh quotient until the secret rate reaches the convergence condition. Finally, perform eigenvalue decomposition to output the optimal antenna phase and amplitude configuration matrix.
[0033] The beneficial effects of this invention are as follows:
[0034] This invention abandons the unrealistic premise of assuming the location of the eavesdropper is known in traditional schemes. It innovatively utilizes the ISAC dual-function beam to achieve a closed-loop chain from radar detection of potential threats to mathematical modeling based on measured data, and finally to precise physical layer shielding. This gives satellite communication systems an active defense capability for the first time. It also endows the security boundary with high dynamism: the radius of the uncertainty region used to delineate the eavesdropper in this invention is not a rigid empirical setting, but is tightly coupled to the real-time sensing quality of the ISAC system. When the sensing signal-to-noise ratio is high, the error circle shrinks to save transmission power; when the sensing environment is harsh, the error circle adaptively expands. This dynamic adaptive capability is a significant improvement over traditional fixed-boundary schemes. This scheme also takes into account the channel angle error of legitimate users caused by the high-speed movement of LEO satellites and the inherent measurement error of radar sensing. By using the maximization of the worst-case confidentiality rate (WCSSR) as the global optimization benchmark, the system is forced to find the optimal solution under the worst-case error superposition state. Therefore, even with unavoidable sensing deviations, this invention can still guarantee a security baseline at the physical layer, and its reliability is significantly better than traditional anti-eavesdropping methods that rely solely on expired channel information. Attached Figure Description
[0035] Figure 1 A schematic diagram of a satellite perception-assisted secure transmission process based on dynamic uncertainty modeling.
[0036] Figure 2 Integrated signal transmission architecture diagram
[0037] Figure 3 Modular structure diagram of a satellite secure transmission system for modeling dynamic uncertainties Detailed Implementation
[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0039] This invention proposes a satellite perception-assisted secure communication method based on dynamic uncertainty modeling, such as... Figure 1 As shown, the method includes:
[0040] S1: Construct an integrated signal transmission architecture.
[0041] Construct an integrated signal transmission architecture, such as Figure 2As shown, the specific received signal models of the communication link and radar detection link are as follows: For the first... For a legitimate user, the downlink communication signal received is represented as follows:
[0042]
[0043] For a potential eavesdropper, the intercepted communication signals are represented as follows:
[0044]
[0045] For a satellite-based radar receiving system, the captured echo signal is represented as:
[0046]
[0047] in, Indicates the satellite's arrival at the [number]th Channel response vectors for each user; and Represented as the first The user and the A user-designed transmit beamforming precoding vector; and This indicates a normalized private signal sent to legitimate users; The total number of legitimate users; For the first The variance of the complex Gaussian white noise at each user terminal is... ; This represents the channel response vector from the satellite to the eavesdropper. The noise at the eavesdropper's end is complex Gaussian white noise with variance . ; This represents the finite impulse response filter vector at the radar receiver. The complex amplitude of the target being detected satisfies ; , This is the array response vector; It is a dual-function transmission signal, and , ; The complex Gaussian white noise at the radar receiver has a variance of... .
[0048] S2: Actively detect and lock onto potential eavesdroppers.
[0049] Satellite-side radar receiving filter The acquired echo signal is processed, and the azimuth, elevation, and time delay of the echo are extracted to calculate the estimated center coordinates of the potential eavesdropper. To quantify the sensing quality, the mathematical formula for calculating the signal-to-noise ratio (SNR) of the radar sensing output is:
[0050]
[0051] in, This indicates the output signal-to-noise ratio of the radar receiver filter; To detect the expected variance of the target's complex amplitude; This is the radar receiver filter vector; This is the covariance matrix constructed based on the target array response vector; For the overall transmitted beamforming matrix; The variance of the received noise at the radar end; This represents the conjugate transpose operation of a matrix.
[0052] S3: Establish a multidimensional channel uncertainty model.
[0053] (1) Establish a legal user channel model for low-Earth orbit satellites that includes Doppler frequency shift and transmission delay, expressed as:
[0054]
[0055] in, This indicates the Doppler shift caused by the high-speed movement of low-orbit satellites; Indicates the first The minimum propagation delay across all paths for a user; For instantaneous time, For carrier frequency; To satisfy the Rice distribution of channel gain, its Rice factor is: The average power is ; This is the array response vector of the uniform planar antenna array at the satellite end;
[0056] (2) Establish a legal user angle uncertainty model based on uniform distribution, expressed as:
[0057]
[0058] in, and The first The true azimuth and elevation angles of each user relative to the satellite antenna array; and The received estimated angle; and The angular deviation caused by channel timeliness is modeled as a random variable following a uniform distribution, i.e. , ;
[0059] (3) Establish a bounded uncertainty region model for the eavesdropper norm based on perception assistance, expressed as:
[0060]
[0061] in, This is the actual array response vector of the eavesdropper; To estimate the response vector of the eavesdropper, which is calculated using radar echoes; This represents the channel estimation error vector; Representing vectors Norm; The radius of the uncertainty region boundary is dynamically calculated based on the perceived signal-to-noise ratio.
[0062] S4: Jointly design the optimal beamforming and receiving filter.
[0063] The specific form of the optimization problem established when jointly designing the optimal beamforming and receiving filter is as follows:
[0064]
[0065]
[0066]
[0067] Among them, in the objective function Indicates the first Worst-case confidentiality rate for each user; To meet the preset radar target detection threshold requirements, ensure that the output signal-to-noise ratio of the detection beam meets the perception baseline; The total transmit power budget for the satellite system is used to constrain the total energy consumption of all dual-function beams;
[0068] Furthermore, the first Worst-case secrecy rate for each user Defined as:
[0069]
[0070] in, For legitimate users achievable data rates For the eavesdropper to intercept the first The achievable rate of a user signal, Representation function Operations are performed to ensure a non-negative secrecy rate.
[0071] S5: Mathematical transformations and iterative algorithm solutions.
[0072] Since the optimization problem established in S4 is highly non-convex, it is extremely difficult to solve directly. The specific operation of handling non-convex optimization problems using semidefinite relaxation and convex-concave processes is as follows:
[0073] For the transmitted beam vector Introducing a positive semidefinite matrix Reduce the dimensionality of the original problem and relax it;
[0074] Addressing the complex non-convex user data rate caused by channel distribution This can be approximated as the difference between two concave functions, i.e. ;in,
[0075]
[0076]
[0077] Introducing auxiliary variables and in the In the next iteration, the subtraction term is subjected to a first-order Taylor expansion. By approximation, we obtain the iterative formula:
[0078]
[0079] in, Represents the trace operation of a matrix; The distribution correlation matrix of the channel from the perspective of legitimate users; and Representing respectively in the The emission covariance matrix variable updated in the next iteration.
[0080] When dealing with the norm-bounded uncertainty region of an eavesdropper, the S-Procedure theorem is used to transform the security constraints containing infinitely many uncertain variables into finite-dimensional linear matrix inequalities (LMIs). The specific transformed constraint matrix form is as follows:
[0081]
[0082] in, The matrix indicates that it is a positive semi-definite matrix; The introduced slack variables satisfy the linear matrix inequality form; For dimensions and number of satellite array antennas Consistent identity matrix; For the first The target beam covariance matrix of the next iteration; Estimate the response vector for radar-detected eavesdroppers; The aggregate scalar that integrates error boundary constraints and noise is specifically expressed as follows:
[0083]
[0084] The specific method for alternately optimizing radar receiver filters is as follows:
[0085] Given the covariance matrices of legitimate users and eavesdroppers Under the premise that the original joint optimization problem is transformed into a typical generalized Rayleigh quotient optimization subproblem, the mathematical model of which is:
[0086]
[0087] By calculating the generalized matrix The maximum eigenvalue is determined, and the corresponding eigenvector is extracted. This eigenvector is then directly used as the updated optimal radar receiving filter vector. The next round of global iteration is carried out, and finally the optimal covariance matrix is decomposed into eigenvalues to output beamforming vectors to control the radio frequency link.
[0088] The above-described embodiments further illustrate the purpose, technical solution, and advantages of the present invention. It should be understood that the above-described embodiments are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made to the present invention within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A satellite-sensing-assisted secure communication method based on dynamic uncertainty modeling, characterized in that, include: S1: Construct an integrated signal transmission architecture; S2: Actively detect and lock onto potential eavesdroppers; S3: Establish a multidimensional channel uncertainty model; S4: Jointly design the optimal beamforming and receiving filter; S5: Mathematical transformations and iterative algorithm solutions.
2. The satellite perception-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, In the S1 integrated signal transmission architecture described above, the specific receiving signal models for the communication link and radar detection link are as follows: For the first... For a legitimate user, the downlink communication signal received is represented as follows: ， For a potential eavesdropper, the intercepted communication signals are represented as follows: ， For a satellite-based radar receiving system, the captured echo signal is represented as: ， in, Indicates the satellite's arrival at the [number]th Channel response vectors for each user; and Represented as the first The user and the first A user-designed transmit beamforming precoding vector; and This indicates a normalized private signal sent to legitimate users; The total number of legitimate users; For the first The variance of the complex Gaussian white noise at each user terminal is... ; This represents the channel response vector from the satellite to the eavesdropper. The noise at the eavesdropper's end is complex Gaussian white noise with variance . ; This represents the finite impulse response filter vector at the radar receiver. The complex amplitude of the target being detected satisfies ; , This is the array response vector; It is a dual-function transmission signal, and , ; The complex Gaussian white noise at the radar receiver has a variance of... .
3. The satellite sensing-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, The process of calculating the sensing signal-to-noise ratio using the signal output from the radar receiving filter to evaluate detection performance includes: The mathematical formula for calculating the signal-to-noise ratio of radar sensing output is: ， in, This indicates the output signal-to-noise ratio of the radar receiver filter; To detect the expected variance of the target's complex amplitude; This is the radar receiver filter vector; This is the covariance matrix constructed based on the target array response vector; For the overall transmitted beamforming matrix; The variance of the received noise at the radar end; This represents the conjugate transpose operation of a matrix.
4. The satellite sensing-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, Establishing a multidimensional channel uncertainty model specifically includes: (1) Establish a legal user channel model for low-Earth orbit satellites that includes Doppler frequency shift and transmission delay, expressed as: ， in, This indicates the Doppler shift caused by the high-speed movement of low-orbit satellites; Indicates the first The minimum propagation delay across all paths of a user; For instantaneous time, For carrier frequency; To satisfy the Rice distribution of channel gain, its Rice factor is: The average power is ; This is the array response vector of the uniform planar antenna array at the satellite end; (2) Establish a legal user angle uncertainty model based on uniform distribution, expressed as: ， in, and The first The true azimuth and elevation angles of each user relative to the satellite antenna array; and The received estimated angle; and The angular deviation caused by channel timeliness is modeled as a random variable following a uniform distribution, i.e. , ; (3) Establish a bounded uncertainty region model for the eavesdropper norm based on perception assistance, expressed as: ， in, This is the actual array response vector of the eavesdropper; To estimate the response vector of the eavesdropper, which is calculated using radar echoes; This represents the channel estimation error vector; Representing vectors Norm; The radius of the uncertainty region boundary is dynamically calculated based on the perceived signal-to-noise ratio.
5. The satellite perception-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, The specific form of the optimization problem established when jointly designing the optimal beamforming and receiving filter is as follows: ， ， ， Among them, in the objective function Indicates the first Worst-case confidentiality rate for each user; To meet the preset radar target detection threshold requirements, ensure that the output signal-to-noise ratio of the detection beam meets the perception baseline; The total transmit power budget for the satellite system is used to constrain the total energy consumption of all dual-function beams; Furthermore, the first Worst-case secrecy rate for each user Defined as: ， in, For legitimate users achievable data rates For the eavesdropper to intercept the first The achievable rate of a user signal, Representation function Operations are performed to ensure a non-negative secrecy rate.
6. The satellite sensing-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, The specific steps for handling non-convex optimization problems using semidefinite relaxation and convex-concave processes are as follows: For the transmitted beam vector Introducing a positive semidefinite matrix Reduce the dimensionality of the original problem and relax it; Addressing the complex non-convex user data rate caused by channel distribution This can be approximated as the difference between two concave functions, i.e. ;in, ， ， Introducing auxiliary variables and in the In the next iteration, the subtraction term is subjected to a first-order Taylor expansion. By approximation, we obtain the iterative formula: ， in, Represents the trace operation of a matrix; The distribution correlation matrix of the channel from the perspective of legitimate users; and They represent the first time in the second month. The emission covariance matrix variable updated in the next iteration.
7. The satellite perception-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, When dealing with the norm-bounded uncertainty region of an eavesdropper, the S-Procedure theorem is used to transform the security constraints containing infinitely many uncertain variables into finite-dimensional linear matrix inequalities (LMIs). The specific transformed constraint matrix form is as follows: ， in, The matrix indicates that it is a positive semi-definite matrix; The introduced slack variables satisfy the linear matrix inequality form; For dimensions and number of satellite array antennas Consistent identity matrix; For the first The target beam covariance matrix of the next iteration; Estimate the response vector for radar-detected eavesdroppers; The aggregate scalar that integrates error boundary constraints and noise is specifically expressed as follows: 。 8. The satellite perception-assisted secure communication method based on dynamic uncertainty modeling according to claim 1, characterized in that, The specific method for alternately optimizing radar receiver filters is as follows: Given the covariance matrices of legitimate users and eavesdroppers Under the premise that the original joint optimization problem is transformed into a typical generalized Rayleigh quotient optimization subproblem, the mathematical model of which is: ， By calculating the generalized matrix The maximum eigenvalue is determined, and the corresponding eigenvector is extracted. This eigenvector is then directly used as the updated optimal radar receiving filter vector. The next round of global iteration is carried out, and finally the optimal covariance matrix is decomposed into eigenvalues to output beamforming vectors to control the radio frequency link.