A Physical Layer Key-Based Secure Communication Method for Unmanned Aerial Vehicles Based on Orthogonal Time-Frequency Space

Abstract

This invention relates to a method for secure physical layer communication for unmanned aerial vehicles (UAVs) based on orthogonal time-frequency space. The method includes the following steps: establishing an UAV air-to-ground communication system consisting of a multi-antenna UAV base station, a legitimate ground-based user with a single antenna, and an unauthorized ground-based user passively eavesdropping; obtaining channel estimation vectors for all dominant paths using a three-dimensional structured orthogonal matching pursuit algorithm; obtaining channel eigenvalues ​​of the communication channel using a path gain adaptive weighted moving average coding method; performing double-bit uniform quantization; and enhancing security through Winnow key negotiation based on dynamic grouping and SHA-2 hash function, thereby obtaining the physical layer key. This invention effectively combats the effects of Doppler shift by transforming the channel dimension using orthogonal time-frequency space technology, improving channel reciprocity, reducing key inconsistency rate, and significantly improving the security and stability of the communication system.

Description

Technical Field

[0001] This invention relates to the field of communication technology, specifically to a method for secure communication of UAV physical layer keys based on orthogonal time-frequency space in air-to-ground communication scenarios. Background Technology

[0002] Unmanned aerial vehicle (UAV) communication, characterized by high link gain, high mobility, and flexible networking, is widely used in both military and civilian fields. In the military, large numbers of UAVs form wide-area distributed, rapid-response communication networks, capable of completing various tasks in complex situations. In the civilian communications field, UAV communication systems can efficiently connect with ground networks, providing small-scale temporary network services in emergencies such as disaster relief, field reconnaissance, and public security when fixed communication infrastructure is unavailable, thus ensuring people's production and daily lives. It is clear that in today's era of rapid technological development, UAV communication has become an indispensable and important form of communication for global public and national defense security.

[0003] However, this technology also exposes some significant vulnerabilities. Due to the broadcast nature of drone wireless transmission and the line-of-sight link characteristic, transmitted information is easily eavesdropped on, severely compromising information transmission security and causing incalculable losses. In a combat environment, the enemy can exploit these vulnerabilities to interfere with or steal battlefield environmental awareness information, collected data, and imagery, which could rapidly alter the battlefield situation. In match verification communications, interference with or theft of account, authentication, and personal information transmitted by base stations could lead to property damage. From the above analysis, we can see that the reliability of drone communication directly impacts the decisive role in the future intelligent unmanned battlefield and the safety of people's lives in multiple fields. Therefore, ensuring reliable and secure drone communication is of paramount importance.

[0004] Due to the inherent characteristics of multi-UAV communication systems, traditional upper-layer encryption technologies are difficult to apply independently, mainly for the following reasons:

[0005] (1) The UAV communication system itself has limited computing power, and the computation time and communication overhead of the authentication key negotiation protocol often impose a large burden on the UAV network.

[0006] (2) The communication mechanism of UAVs is complex, information exchange is frequent, and the network scale and structure are easily changed, which makes it difficult to distribute, share and manage keys online;

[0007] (3) The standardized protection in the UAV communication network is not secure enough and is very vulnerable to attack, which reduces security.

[0008] Unlike upper-layer encryption technologies, physical layer security technologies focus on utilizing the characteristics of physical media and communication devices to achieve secure information transmission. Its goal is to prevent information leakage and unauthorized access at the physical level without relying on traditional encryption algorithms. Among these technologies, physical layer key generation based on the reciprocity mechanism of wireless channels has been widely studied both domestically and internationally due to its lightweight and flexible advantages.

[0009] However, unlike traditional terrestrial wireless communication systems, UAV air-to-ground communication systems are characterized by a large relative movement speed between the communication terminal and the UAV base station, which will produce a serious Doppler frequency shift. The traditional time-frequency domain channel estimation results have a large deviation, making it difficult for legitimate communication parties to generate a consistent key, which seriously affects the key generation performance. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a physical layer key-based secure communication method for UAVs based on orthogonal time-frequency space. By converting the rapidly changing time-frequency domain channel into a slowly changing time-delay Doppler domain channel, and then extracting channel feature values ​​to generate physical layer keys, this method effectively addresses the challenges brought about by channel changes in the high-speed moving environment of UAVs, ensuring the stability and security of communication.

[0011] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a method for secure communication of UAV physical layer keys based on orthogonal time-frequency space, comprising the following steps:

[0012] Step 1: Establish an UAV air-to-ground communication system consisting of a multi-antenna UAV base station, a single-antenna authorized ground user authorized to receive information, and an unauthorized ground user passively eavesdropping. Then, perform time-delay Doppler angle dimension channel modeling on the channel of the UAV air-to-ground communication system to obtain the time-delay Doppler domain channel matrix.

[0013] Step 2: Using a three-dimensional structured orthogonal matching pursuit algorithm, based on the communication channel between a multi-antenna UAV base station and a single-antenna legitimate user, the index set of the dominant path of the communication channel in the delay dimension, Doppler dimension, and angle dimension is obtained. The channel estimation vector of all dominant paths is determined by the least squares method.

[0014] Step 3: Adaptively weight the path gain strength of the channel estimation vector of the dominant path to obtain the channel characteristic value of the communication channel, which is used to improve the reciprocity of the communication channel between the multi-antenna UAV base station and the single-antenna legitimate user, and ensure that the two communicating parties can generate consistent physical layer keys.

[0015] Step 4: Perform double-bit uniform quantization on the channel feature values ​​of the communication channel to generate a channel feature value sequence, and then encode the channel feature value sequence to obtain the initialized bit sequence.

[0016] Step 5: Perform Winnow key negotiation based on dynamic grouping on the initialized bit sequence. This means that a consistent physical layer key for encrypting the transmitted information is obtained between the multi-antenna UAV base station and the single-antenna legitimate user. Encryption using the physical layer key increases the bit error rate of unauthorized users, making it impossible for them to eavesdrop, and achieving secure communication using the physical layer key.

[0017] Furthermore, step one specifically includes the following steps:

[0018] Step 11: First, model the channel of the UAV air-to-ground communication system. Since the downlink channel... The channel estimation method is the same for all single-antenna legitimate users, so channel estimation is only performed for any single-antenna legitimate user. The specific steps are as follows:

[0019] Definition and the first of multi-antenna UAV base stations The correlated time-domain time-varying channel of each antenna for:

[0020] ,

[0021] In the formula, For time-domain discretization indexes, For symbol period, Subcarrier spacing; The sampling interval is... The number of sampling points within the symbol period. An index for time-delay domain discretization;

[0022] Number of main paths in time-domain time-varying channels Each main path contains Subpath, the first The Doppler frequency shift of each subpath is , For the flight speed of multi-antenna drone base stations, For carrier frequency, The speed of light; the channel gain is expressed as , No. The latency of all sub-paths under the main path is 1. Antenna quantity index , This represents the total number of antennas in a multi-antenna drone base station.

[0023] No. Antenna launch angle of each sub-path for Considering a linear uniform antenna array, and The relevant spatial angle is , The distance between the antennas. The wavelength of the subcarrier;

[0024] This is the response of the pulse shaping filter;

[0025] Then we get the first Equivalent Channel Response Matrix in Delay-Doppler Domain on a Single Antenna for:

[0026] ,

[0027] In the formula, variables , For the first Doppler frequency shift of each sub-path The number of discrete symbols in the time delay domain. Symbolic index for the Doppler field, Symbolic index for time-delay discrete domain right Take the mold. For the first The latency of all sub-paths under the main path. For the first The antenna spatial angle of each subpath, with respect to variables function for:

[0028] ,

[0029] In the formula, The number of symbols in the Doppler field. For summation index variables;

[0030] Step 12: In order to perform sparse representation of the channel of the UAV air-to-ground communication system, the equivalent channel response matrix in the time-delay Doppler domain is calculated along the spatial domain. Performing the inverse discrete Fourier transform yields the time-delay Doppler angle-dimensional channel. :

[0031] ,

[0032] In the formula, For angle index, This represents the total number of antennas in a multi-antenna drone base station.

[0033] Step 13: In the time-delay Doppler angle dimension channel After modeling, the received signal of a legitimate user with a single antenna With time-delay Doppler domain channel matrix They are represented as follows:

[0034] ,

[0035] ,

[0036] In the formula, Pilot matrix for drone base station transmission, for The identity matrix, The number of drone base station antennas. It is a sparse basis matrix. It is additive white Gaussian noise. It represents the Kronecker product.

[0037] Furthermore, step two specifically includes the following steps:

[0038] Step 21: Set the loop count index for the 3D structured orthogonal matching pursuit algorithm. Dominant path index set of the time-delay Doppler domain channel matrix The channel estimation vector of the initial time-delay Doppler domain channel matrix Initial residual for:

[0039] ,

[0040] In the formula, It is the received signal of a legitimate user with a single antenna. It is the time-delay Doppler domain channel matrix;

[0041] Step 22: Based on the three-dimensional structured orthogonal matching pursuit algorithm, iterate cyclically according to the number of paths to be estimated, and then calculate the correlation vector of the time-delay Doppler domain channel matrix. for:

[0042] ,

[0043] In the formula, The conjugate transpose of the time-delay Doppler domain channel matrix. For the first The residual of the next iteration;

[0044] Furthermore, the relevant vectors Rearranged into a three-dimensional tensor with delay, Doppler, and antenna dimensions. for:

[0045] ,

[0046] In the formula, The length of the protection interval in the delay dimension. The length of the protection interval in the Doppler dimension. The number of antennas for a multi-antenna drone base station;

[0047] Step 23: Calculate the three-dimensional tensor Expanding along the delay dimension yields the expanded matrix. for:

[0048] ,

[0049] Calculate the row vectors of the expanded matrix. Norm, to obtain the correlation vector of the delay dimension Select The delay dimension index corresponding to the maximum value in the middle ,Then As the first The support set for the next iteration in the delay dimension;

[0050] Step 24: Fixed Delay Dimension Index The first slice matrix of the expanded matrix is:

[0051] ,

[0052] Then, the row vectors of the first slice matrix along the Doppler dimension are calculated. Norm, to obtain the Doppler correlation vector By selecting the maximum value in the Doppler correlation vector, the non-zero block in the Doppler dimension is determined, and thus the first... The support set of the next iteration in the Doppler dimension ;

[0053] Step 25: Based on the first slice matrix and the support set in the Doppler dimension The second slice matrix is ​​obtained as follows:

[0054] ,

[0055] Calculate the column vectors of the second slice matrix along the angular dimension. Norm, to obtain the angular dimension correlation vector ;

[0056] Furthermore, by utilizing the lifting transformation method, the angle-dimensional correlation vector is... Transformation and processing are performed to determine the first... The support set of the next iteration in the angular dimension ;

[0057] Step 26: Support the set of delay dimension, Doppler dimension, and angle dimension determined in steps 23 to 25 and update it to obtain the index set. for:

[0058] ,

[0059] Therefore, the channel estimation vector is obtained by using the least squares method. for:

[0060] ,

[0061] Next, update the residual to This is used to eliminate the influence of the estimated dominant paths on the residuals. Steps 23 to 26 are repeated until all dominant paths have been detected and estimated. Finally, the channel estimation vector of the time-delay Doppler domain channel matrix is ​​output. for:

[0062] ,

[0063] In the formula, The number of dominant paths.

[0064] Furthermore, step three specifically includes the following steps:

[0065] Step 31, the channel estimation vector is denoted as:

[0066] ,

[0067] In the formula, The number of dominant paths;

[0068] Step 32: Then, based on the magnitude of the dominant path gain strength, sort the dominant path gain strengths from smallest to largest to obtain:

[0069] ;

[0070] In the formula, This is an estimate of the dominant path with the smallest gain intensity after sorting. For the sorted number Gain strength estimates for each dominant path, This is the estimated value of the dominant path with the largest gain after sorting;

[0071] Step 33: Adaptively weight the dominant path gain strength from 0 to 1 after sorting, i.e.:

[0072] ,

[0073] In the formula, For the first Each weight is used to adaptively weight the gain strength of the ranked dominant paths, and its value ranges from 0 to 1. The path corresponding to the minimum gain. The intermediate weight corresponding to the path with the maximum gain Increasing linearly;

[0074] Step 34: Restore to the original order, perform sliding windowing processing, and obtain the channel characteristic values ​​of the communication channel. :

[0075] ,

[0076] In the formula, For window size, The number of channel eigenvalues. This represents the overlap size between two adjacent windows. For the first The size of the adaptive weighted weights, Indicates the first An adaptively weighted estimate of the dominant path gain strength.

[0077] Furthermore, step four specifically includes the following steps:

[0078] Step 41: The channel feature value sequences for multi-antenna UAV base stations and single-antenna legitimate users are as follows: and The sample size is ;

[0079] Then, the multi-antenna UAV base station calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ;

[0080] A single-antenna legitimate user calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ;

[0081] Based on the number of quantization bits Calculate the number of quantized regions Next, the sample intervals are divided into equal intervals as follows:

[0082] ,

[0083] ,

[0084] In the formula, and The classification intervals are for multi-antenna drone base stations and single-antenna legitimate users, respectively.

[0085] Step 42: Multi-antenna drone base stations and single-antenna legitimate users are classified by interval. and Calculate the quantization threshold, and then use the quantization threshold to quantize the sequence. and In Each sample was classified separately.

[0086] Step 43: The multi-antenna UAV base station and the single-antenna legitimate user encode the classified samples using Gray code to obtain the initialized bit sequence. and .

[0087] Furthermore, step five specifically includes the following steps:

[0088] Step 51: The multi-antenna drone base station and the single-antenna legitimate user will initialize the bit sequence. and Calculate the initial bit sequence by randomly permuting it according to the same rules. and Bit inconsistency rate And set a bit inconsistency rate threshold. ;

[0089] like Then the initialized bit sequence after random permutation and according to Length grouping, Indicates the initial bit sequence Length;

[0090] like Then the initialized bit sequence after random permutation and according to Length groups are formed, parity bits for each group are calculated, and the groups are exchanged on a common channel.

[0091] Step 52: Compare the parity check bits exchanged between the multi-antenna UAV base station and the single-antenna legitimate user. If any of the parity check bits are different, perform error correction only on the packets with different parity check bits, discard the first bit of the packet, and use the remaining bits to form a Hamming code group. and Perform local error correction;

[0092] Step 53: Calculate the Hamming code group and The accompanying , They are respectively:

[0093] , ,

[0094] In the formula, For the verification matrix, This is the conjugate transpose of the parity check matrix;

[0095] Step 54: The multi-antenna UAV base station transmits the associated signal through a public channel. Send to a single-antenna legitimate user, who then performs a bitwise XOR calculation. ,in The sign is calculated by XORing, thereby obtaining the position of the erroneous bit and realizing a round of dynamic negotiation and error correction;

[0096] Step 55: Repeat steps 51 to 54 until the initial keys of the multi-antenna drone base station and the single-antenna legitimate user are completely identical, thus obtaining the physical layer key. .

[0097] Furthermore, in order to improve the randomness and security of the physical layer key, the method also includes a step of performing SHA-2 hash function encryption on the negotiated physical layer key.

[0098] Furthermore, the specific steps for enhancing the confidentiality of the negotiated physical layer key using the SHA-2 hash function are as follows:

[0099] Step 61: Based on the SHA-2 algorithm, the initial physical layer key is consistent between the multi-antenna UAV base station and the single-antenna legitimate user after key negotiation. Fill in the data, and then set an initial hash value, which consists of eight 32-bit fixed constants;

[0100] Step 62: The multi-antenna drone base station and the single-antenna legitimate user further divide the message into message blocks, and then perform expansion and round-iteration operations on each message block. In each round, logical functions and bit operations are used to gradually integrate the message block information into the intermediate hash value.

[0101] Step 63: After processing all message blocks, add the initial hash value and the intermediate results during processing to each hash value in 32-bit increments to produce a final 256-bit hash value, which is the physical layer key enhanced with the SHA-2 hash function. .

[0102] The beneficial effects of this invention are as follows: This invention proposes a physical layer key-secured communication method for UAVs based on orthogonal time-frequency space. It uses a three-dimensional structured orthogonal matching pursuit algorithm to estimate the communication channel between a multi-antenna UAV base station and a single-antenna legitimate user. Then, through the proposed path gain adaptive weighted moving average coding algorithm, double-bit uniform quantization, Winnow key negotiation based on dynamic grouping, and hash function-based security enhancement, a physical layer key is obtained for encryption.

[0103] Simulation results show that the physical layer key-secured communication scheme proposed in this invention can effectively combat the Doppler shift problem, reduce the impact of channel changes on communication performance, improve legitimate channel reciprocity, reduce key inconsistency rate, and ensure the security and stability of the communication system. Attached Figure Description

[0104] Figure 1 This is a framework diagram of the UAV air-to-ground communication system of the present invention;

[0105] Figure 2 This is a schematic diagram of the time-delay Doppler angle dimension channel of the UAV air-to-ground communication system of the present invention;

[0106] Figure 3 This is a diagram showing the channel reciprocity results of the preprocessing algorithm proposed in the simulation experiment of this invention under different signal-to-noise ratios;

[0107] Figure 4 This is a graph showing the key inconsistency rate results of the preprocessing algorithm proposed in the simulation experiment of this invention under different signal-to-noise ratios;

[0108] Figure 5 This is a bit error rate diagram of different physical layer security schemes under different signal-to-noise ratios in the simulation experiment of this invention. Detailed Implementation

[0109] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0110] To achieve the above objectives, the present invention provides the following specific embodiments:

[0111] Example 1: As Figure 1 , Figure 2 As shown, a method for secure physical layer key communication for unmanned aerial vehicles based on orthogonal time-frequency space includes the following steps:

[0112] S01, such as Figure 1 As shown, a multi-antenna drone base station is established. A UAV air-to-ground communication system is composed of a single-antenna legitimate user authorized to receive information on the ground and a ground-based unauthorized user passively eavesdropping. This allows for time-delay Doppler angle-dimensional channel modeling of the UAV air-to-ground communication system channel.

[0113] S02, due to the downlink channel The channel estimation method is the same for all single-antenna legitimate users, so channel estimation is only performed for any single-antenna legitimate user:

[0114] Definition and the first of multi-antenna UAV base stations The correlated time-domain time-varying channel of each antenna for:

[0115] ,

[0116] In the formula, For time-domain discretization indexes, For symbol period, Subcarrier spacing; The sampling interval is... The number of sampling points within the symbol period. An index for time-delay domain discretization;

[0117] Number of main paths in time-domain time-varying channels Each main path contains Subpath, the first The Doppler frequency shift of each subpath is , For the flight speed of multi-antenna drone base stations, For carrier frequency, The speed of light; the channel gain is expressed as , No. The latency of all sub-paths under the main path is 1. Antenna quantity index , This represents the total number of antennas in a multi-antenna drone base station.

[0118] No. Antenna launch angle of each sub-path for Considering a linear uniform antenna array, and The relevant spatial angle is , The distance between the antennas. The wavelength of the subcarrier;

[0119] Given the pulse shaping filter response; then we obtain the first... Equivalent Channel Response Matrix in Delay-Doppler Domain on a Single Antenna for:

[0120] ,

[0121] In the formula, variables , For the first Doppler frequency shift of each sub-path The number of discrete symbols in the time delay domain. Symbolic index for the Doppler field, Symbolic index for time-delay discrete domain right Take the mold. For the first The latency of all sub-paths under the main path. For the first The antenna spatial angle of each subpath, with respect to variables function for:

[0122] ,

[0123] In the formula, The number of symbols in the Doppler field. For summation index variables.

[0124] S03. In order to sparsely represent the channel of the UAV air-to-ground communication system, the equivalent channel response matrix in the time-delay Doppler domain is obtained along the spatial domain. Performing the inverse discrete Fourier transform yields the time-delay Doppler angle-dimensional channel. :

[0125] ,

[0126] In the formula, For angle index, This represents the total number of antennas in a multi-antenna drone base station.

[0127] S04, In the time delay Doppler angle dimension channel After modeling, the received signal of a legitimate user with a single antenna is obtained. With time-delay Doppler domain channel matrix , respectively represented as:

[0128] ,

[0129] ,

[0130] In the formula, Pilot matrix for drone base station transmission, for The identity matrix, The number of drone base station antennas. It is a sparse basis matrix. It is additive white Gaussian noise. It represents the Kronecker product.

[0131] S05. Employ a three-dimensional structured orthogonal matching pursuit algorithm, based on the communication channel between a multi-antenna UAV base station and a single-antenna legitimate user, and set the loop count index for the three-dimensional structured orthogonal matching pursuit algorithm. Dominant path index set of the time-delay Doppler domain channel matrix The channel estimation vector of the initial time-delay Doppler domain channel matrix Initial residual for:

[0132] ,

[0133] In the formula, It is the received signal of a legitimate user with a single antenna. It is the time-delay Doppler domain channel matrix.

[0134] S06. Iterate cyclically based on the number of paths to be estimated, and then calculate the correlation vector of the time-delay Doppler domain channel matrix. for:

[0135] ,

[0136] In the formula, The conjugate transpose of the time-delay Doppler domain channel matrix. For the first The residual of the next iteration;

[0137] Furthermore, the relevant vectors Rearranged into a three-dimensional tensor with delay, Doppler, and antenna dimensions. for:

[0138] ,

[0139] In the formula, The length of the protection interval in the delay dimension. The length of the protection interval in the Doppler dimension. This refers to the number of antennas in a multi-antenna drone base station.

[0140] S07, Calculating 3D Tensors Expanding along the delay dimension yields the expanded matrix. for:

[0141] ,

[0142] Calculate the row vectors of the expanded matrix. Norm, to obtain the correlation vector of the delay dimension Select The delay dimension index corresponding to the maximum value in the middle ,Then As the first The support set in the delay dimension of the next iteration.

[0143] S08, Fixed Delay Dimension Index The first slice matrix of the expanded matrix is:

[0144] ,

[0145] Then, the row vectors of the first slice matrix along the Doppler dimension are calculated. Norm, to obtain the Doppler correlation vector By selecting the maximum value in the Doppler correlation vector, the non-zero block in the Doppler dimension is determined, and thus the first... The support set of the next iteration in the Doppler dimension .

[0146] S09, Based on the first slice matrix and the support set in the Doppler dimension The second slice matrix is ​​obtained as follows:

[0147] ,

[0148] Calculate the column vectors of the second slice matrix along the angular dimension. Norm, to obtain the angular dimension correlation vector ;

[0149] Furthermore, by utilizing the lifting transformation method, the angle-dimensional correlation vector is... Transformation and processing are performed to determine the first... The support set of the next iteration in the angular dimension .

[0150] S10. Support the set and update the delay dimension, Doppler dimension and angle dimension determined in S07 to S09 to obtain the index set. for:

[0151] ,

[0152] Therefore, the channel estimation vector is obtained by using the least squares method. for:

[0153] ,

[0154] Next, update the residual to This is used to eliminate the influence of the estimated dominant paths on the residuals. Steps 23 to 26 are repeated until all dominant paths have been detected and estimated. Finally, the channel estimation vector of the time-delay Doppler domain channel matrix is ​​output. for:

[0155] ,

[0156] In the formula, The number of dominant paths.

[0157] S11, the channel estimation vector is denoted as:

[0158] ,

[0159] In the formula, The number of dominant paths;

[0160] S12. Then, based on the magnitude of the dominant path gain strength, sort the dominant path gain strengths from smallest to largest to obtain:

[0161] ;

[0162] In the formula, This is an estimate of the dominant path with the smallest gain intensity after sorting. For the sorted number Gain strength estimates for each dominant path, This is the estimated value of the dominant path with the largest gain after sorting;

[0163] S13. Adaptively weight the gain intensity of the dominant path after sorting from 0 to 1, i.e.:

[0164] ,

[0165] In the formula, For the first Each weight is used to adaptively weight the gain strength of the ranked dominant paths, and its value ranges from 0 to 1. The path corresponding to the minimum gain. The intermediate weight corresponding to the path with the maximum gain Increasing linearly;

[0166] S14. Restore to the original order, perform sliding windowing processing, and obtain the channel characteristic values ​​of the communication channel. :

[0167] ,

[0168] In the formula, For window size, The number of channel eigenvalues. This represents the overlap size between two adjacent windows. For the first The size of the adaptive weighted weights, Indicates the first An adaptively weighted estimate of the dominant path gain strength.

[0169] S15. Perform double-bit uniform quantization on the channel characteristic values ​​of the communication channel. The channel characteristic value sequences for multi-antenna UAV base stations and single-antenna legitimate users are as follows: and The sample size is ;

[0170] Then, the multi-antenna UAV base station calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ;

[0171] A single-antenna legitimate user calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ;

[0172] Based on the number of quantization bits Calculate the number of quantized regions Next, the sample intervals are divided into equal intervals as follows:

[0173] ,

[0174] ,

[0175] In the formula, and The classification intervals are for multi-antenna drone base stations and single-antenna legitimate users, respectively.

[0176] S16, multi-antenna drone base stations and single-antenna legitimate users are classified by interval and Calculate the quantization threshold, and then use the quantization threshold to quantize the sequence. and In Each sample was classified separately.

[0177] S17. Multi-antenna UAV base stations and single-antenna legitimate users encode the classified samples using Gray codes to obtain an initialized bit sequence. and .

[0178] S18, multi-antenna drone base stations and single-antenna legitimate users will initialize the bit sequence and Calculate the initial bit sequence by randomly permuting it according to the same rules. and Bit inconsistency rate And set a bit inconsistency rate threshold. ;

[0179] like Then the initialized bit sequence after random permutation and according to Length grouping, Indicates the initial bit sequence Length;

[0180] like Then the initialized bit sequence after random permutation and according to The length is grouped, the parity bit of each group is calculated, and the groups are exchanged on the common channel.

[0181] S19. Compare the parity check bits exchanged between the multi-antenna UAV base station and the single-antenna legitimate user. If any of the parity check bits are different, then only the parity check bits that are different are corrected. The first bit of the different parity check bits is discarded, and the remaining bits form a Hamming code group. With Local error correction.

[0182] S20. Calculate the Hamming code group. and The accompanying , They are respectively:

[0183] , ,

[0184] In the formula, For the verification matrix, It is the conjugate transpose of the parity matrix.

[0185] S21, multi-antenna drone base station transmits accompanying signals through open channels Send to a single-antenna legitimate user, who then performs a bitwise XOR calculation. ,in The sign is calculated by XORing, thus obtaining the position of the erroneous bit, and realizing a round of dynamic negotiation and error correction.

[0186] S22. Repeat S18 to S21 until the initial keys of the multi-antenna drone base station and the single-antenna legitimate user are completely identical, thus obtaining the physical layer key. .

[0187] Example 2: Same as Example 1, except that: to improve the randomness and security of the physical layer key, it also includes a step of performing SHA-2 hash function encryption on the negotiated physical layer key, specifically:

[0188] S23. Based on the SHA-2 algorithm, after key negotiation, the multi-antenna UAV base station and the single-antenna legitimate user will agree on the initial physical layer key. Fill in the data and then set an initial hash value, which consists of eight 32-bit fixed constants.

[0189] S24. The multi-antenna drone base station and the single-antenna legitimate user then divide the message into message blocks, and then perform expansion and round-by-round iterative operations on each message block. In each round, logical functions and bit operations are used to gradually integrate the message block information into the intermediate hash value.

[0190] S25. After processing all message blocks, add the initial hash value and the intermediate results during processing to each hash value in 32-bit increments to produce a final 256-bit hash value, which is the physical layer key after SHA-2 encryption. .

[0191] like Figures 1 to 5 As shown, to further illustrate the technical solution and technical effects of the present invention, the following specific examples are provided:

[0192] Specific example 1: such as Figure 1 , Figure 2 As shown, a method for secure physical layer key communication for unmanned aerial vehicles based on orthogonal time-frequency space includes the following steps:

[0193] S01, such as Figure 1 As shown, a UAV air-to-ground communication system is established, consisting of a multi-antenna UAV base station equipped with 32 uniform antennas, four authorized single-antenna legitimate users for ground-based information reception, and one unauthorized user for ground-based passive eavesdropping. The channel of the UAV air-to-ground communication system is then modeled in time-delay Doppler angle dimension.

[0194] S02. Since the channel estimation method is the same for all four single-antenna legitimate users in the downlink channel, channel estimation is only performed for any one single-antenna legitimate user:

[0195] Definition and the first of multi-antenna UAV base stations The correlated time-domain time-varying channel of each antenna for:

[0196] ,

[0197] In the formula, For time-domain discretization indexes, symbol period Subcarrier spacing Sampling interval , The number of sampling points within the symbol period. An index for time-delay domain discretization;

[0198] Number of main paths in time-domain time-varying channels Each main path contains Subpath, the first The Doppler frequency shift of each subpath is , For the flight speed of multi-antenna drone base stations, For carrier frequency, The speed of light; the channel gain is expressed as , No. The latency of all sub-paths under the main path is 1. , Antenna quantity index , This represents the total number of antennas in a multi-antenna drone base station.

[0199] No. Antenna launch angle of each sub-path for , Considering a linear uniform antenna array, and The relevant spatial angle is , The distance between the antennas. The wavelength of the subcarrier. ;

[0200] This is the response of the pulse shaping filter;

[0201] Then we get the first Equivalent Channel Response Matrix in Delay-Doppler Domain on a Single Antenna for:

[0202] ,

[0203] In the formula, variables , For the first Doppler frequency shift of each sub-path The number of discrete symbols in the time delay domain. Symbolic index for the Doppler field, Symbolic index for time-delay discrete domain right Take the mold. For the first The latency of all sub-paths under the main path. For the first The antenna spatial angle of each subpath, with respect to variables function for:

[0204] ,

[0205] In the formula, The number of symbols in the Doppler field. For summation index variables.

[0206] S03. In order to sparsely represent the channel of the UAV air-to-ground communication system, the equivalent channel response matrix in the time-delay Doppler domain is obtained along the spatial domain. Performing the inverse discrete Fourier transform yields the time-delay Doppler angle-dimensional channel. :

[0207] ,

[0208] In the formula, For angle index, This represents the total number of antennas in a multi-antenna drone base station. For example... Figure 2 As shown, only when , and hour, Only then will there be non-zero elements.

[0209] S04, In the time delay Doppler angle dimension channel After modeling, the received signal of a legitimate user with a single antenna is obtained. With time-delay Doppler domain channel matrix , respectively represented as:

[0210] ,

[0211] ,

[0212] In the formula, Pilot matrix for drone base station transmission, for The identity matrix, The number of drone base station antennas. It is a sparse basis matrix. It is additive white Gaussian noise. It represents the Kronecker product.

[0213] S05. Employ a three-dimensional structured orthogonal matching pursuit algorithm, based on the communication channel between a multi-antenna UAV base station and a single-antenna legitimate user, and set the loop count index for the three-dimensional structured orthogonal matching pursuit algorithm. Dominant path index set of the time-delay Doppler domain channel matrix The channel estimation vector of the initial time-delay Doppler domain channel matrix Initial residual for:

[0214] ,

[0215] In the formula, It is the received signal of a legitimate user with a single antenna. It is the time-delay Doppler domain channel matrix.

[0216] S06. Iterate cyclically based on the number of paths to be estimated, and then calculate the correlation vector of the time-delay Doppler domain channel matrix. for:

[0217] ,

[0218] In the formula, The conjugate transpose of the time-delay Doppler domain channel matrix. For the first The residual of the next iteration;

[0219] Furthermore, the relevant vectors Rearranged into a three-dimensional tensor with delay, Doppler, and antenna dimensions. for:

[0220] ,

[0221] In the formula, The length of the protection interval in the delay dimension. The length of the protection interval in the Doppler dimension. This refers to the number of antennas in a multi-antenna drone base station.

[0222] S07, Calculating 3D Tensors Expanding along the delay dimension yields the expanded matrix. for:

[0223] ,

[0224] Calculate the row vectors of the expanded matrix. Norm, to obtain the correlation vector of the delay dimension Select The delay dimension index corresponding to the maximum value in the middle ,Then As the first The support set in the delay dimension of the next iteration.

[0225] S08, Fixed Delay Dimension Index The first slice matrix of the expanded matrix is:

[0226] ,

[0227] Then, the row vectors of the first slice matrix along the Doppler dimension are calculated. Norm, to obtain the Doppler correlation vector By selecting the maximum value in the Doppler correlation vector, the non-zero block in the Doppler dimension is determined, and thus the first... The support set of the next iteration in the Doppler dimension .

[0228] S09, Based on the first slice matrix and the support set in the Doppler dimension The second slice matrix is ​​obtained as follows:

[0229] ,

[0230] Calculate the column vectors of the second slice matrix along the angular dimension. Norm, to obtain the angular dimension correlation vector ;

[0231] Furthermore, by utilizing the lifting transformation method, the angle-dimensional correlation vector is... Transformation and processing are performed to determine the first... The support set of the next iteration in the angular dimension .

[0232] S10. Support the set and update the delay dimension, Doppler dimension and angle dimension determined in S07 to S09 to obtain the index set. for:

[0233] ,

[0234] Therefore, the channel estimation vector is obtained by using the least squares method. for:

[0235] ,

[0236] Next, update the residual to This is used to eliminate the influence of the estimated dominant paths on the residuals. Steps 23 to 26 are repeated until all dominant paths have been detected and estimated. Finally, the channel estimation vector of the time-delay Doppler domain channel matrix is ​​output. for:

[0237] ,

[0238] In the formula, The number of dominant paths.

[0239] S11, the channel estimation vector is denoted as:

[0240] ,

[0241] In the formula, The number of dominant paths.

[0242] S12. Then, based on the magnitude of the dominant path gain strength, sort the dominant path gain strengths from smallest to largest to obtain:

[0243]

[0244] In the formula, This is an estimate of the dominant path with the smallest gain intensity after sorting. For the sorted number Gain strength estimates for each dominant path, This is the estimated value of the dominant path with the largest gain intensity after sorting.

[0245] S13. Using an adaptive weighting method, the gain intensity of the dominant path is adaptively weighted from 0 to 1, i.e.:

[0246] ,

[0247] In the formula, For the first Each weight is used to adaptively weight the gain strength of the ranked dominant paths, and its value ranges from 0 to 1. The path corresponding to the minimum gain. The intermediate weight corresponding to the path with the maximum gain It increases according to a linear law.

[0248] S14. Restore the original order and perform sliding windowing to obtain the channel characteristic values ​​of the communication channel. for:

[0249] ,

[0250] In the formula, The number of channel features, and the window size. The overlap size of two adjacent windows , For the first The size of the adaptive weighted weights, Indicates the first An adaptively weighted estimate of the dominant path gain strength.

[0251] S15. Perform double-bit uniform quantization on the channel characteristic values ​​of the communication channel. The channel characteristic value sequences for multi-antenna UAV base stations and single-antenna legitimate users are as follows: and The sample size is ;

[0252] Then, the multi-antenna UAV base station calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ;

[0253] A single-antenna legitimate user calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ;

[0254] Based on the number of quantization bits Calculate the number of quantized regions Next, the sample intervals are divided into equal intervals as follows:

[0255] ,

[0256] ,

[0257] In the formula, and The classification intervals are for multi-antenna drone base stations and single-antenna legitimate users, respectively.

[0258] S16, multi-antenna drone base stations and single-antenna legitimate users are classified by interval and Calculate the quantization threshold, and then use the quantization threshold to quantize the sequence. and In Each sample was classified separately.

[0259] S17. Multi-antenna UAV base stations and single-antenna legitimate users encode the classified samples using Gray codes to obtain an initialized bit sequence. and .

[0260] S18, Step 51: The multi-antenna drone base station and the single-antenna legitimate user will initialize the bit sequence. and Calculate the initial bit sequence by randomly permuting it according to the same rules. and Bit inconsistency rate And set a bit inconsistency rate threshold. ;

[0261] like Then the initialized bit sequence after random permutation and according to Length grouping, Indicates the initial bit sequence Length;

[0262] like Then the initialized bit sequence after random permutation and according to The length is grouped, the parity bit of each group is calculated, and the groups are exchanged on the common channel.

[0263] S19. Compare the parity check bits exchanged between the multi-antenna UAV base station and the single-antenna legitimate user. If any parity check bit is different, then only perform error correction operation on the differing packet: discard the first bit of the packet, and the remaining bits form a Hamming code group. and Perform local error correction.

[0264] S20. Calculate the Hamming code group. and The accompanying , They are respectively:

[0265] , ,

[0266] In the formula, For the verification matrix, It is the conjugate transpose of the parity matrix.

[0267] S21, multi-antenna drone base station transmits accompanying signals through open channels Send to a single-antenna legitimate user, who then performs a bitwise XOR calculation. ,in The sign is calculated by XORing, thus obtaining the position of the erroneous bit, and realizing a round of dynamic negotiation and error correction.

[0268] S22. Repeat S18 to S21 until the initial keys of the multi-antenna drone base station and the single-antenna legitimate user are completely identical, thus obtaining the physical layer key. .

[0269] Specific Implementation Example 2: As shown in the example Figures 3 to 5 As shown, the method is the same as in Specific Embodiment 1, except that: in order to improve the randomness and security of the physical layer key, a step of performing SHA-2 hash function encryption on the negotiated physical layer key is also included, specifically:

[0270] S23. Based on the SHA-2 algorithm, after key negotiation, the multi-antenna UAV base station and the single-antenna legitimate user will agree on the initial physical layer key. Fill in the data and then set an initial hash value, which consists of eight 32-bit fixed constants.

[0271] S24. The multi-antenna drone base station and the single-antenna legitimate user then divide the message into message blocks, and then perform expansion and round-by-round iterative operations on each message block. In each round, logical functions and bit operations are used to gradually integrate the message block information into the intermediate hash value.

[0272] S25. After processing all message blocks, add the initial hash value and the intermediate results during processing to each hash value in 32-bit increments to produce a final 256-bit hash value, which is the physical layer key after SHA-2 encryption. .

[0273] Figure 3 The performance of different preprocessing algorithms in improving channel reciprocity is presented, and channel correlation is evaluated using the Pearson correlation coefficient; a higher coefficient indicates better key consistency. As the signal-to-noise ratio (SNR) gradually increases, the channel eigenvalue correlation coefficient gradually rises, and the correlation coefficient of the Moving Average Coding (MAC) algorithm is lower than that of the Adaptive Weighted of Path Gain Moving Average Coding (AWPMAC) algorithm. This is because smaller channel responses are more susceptible to noise in the delay-Doppler domain. By sorting the channel response strengths and adaptively weighting them, the impact of noise is further reduced, thereby improving the reciprocity of legitimate channels and making it easier for multi-antenna UAV base stations and single-antenna legitimate users to generate consistent keys.

[0274] Figure 4 The key inconsistency rate is presented under different preprocessing algorithms. The key inconsistency rate is the ratio of the number of different bits in the quantized keys of legitimate communicating parties to the key length. A smaller value is more conducive to generating a consistent key. As the signal-to-noise ratio (SNR) gradually increases, the key inconsistency rate gradually decreases. The key inconsistency rate of the MAC algorithm is lower than that of the unpreprocessed key inconsistency rate and significantly higher than that of the AWPMAC algorithm. This is because the AWPMAC algorithm improves channel reciprocity. Through a more refined channel feature extraction and processing mechanism, it enables multi-antenna UAV base stations and single-antenna legitimate users to more accurately map channel information into binary bit sequences, reducing quantization errors caused by channel differences.

[0275] Figure 5 The communication bit error rates of the key generation scheme and the artificial noise scheme were compared. In the key generation scheme, the bit error rate in legitimate communication is low. This is because the unauthorized user passively eavesdropping on the ground does not possess the same physical layer key as the sender; it cannot obtain the correct transmission symbols, thus the bit error rate remains at 0.45. In the artificial noise security scheme, although the bit error rate of the unauthorized user passively eavesdropping on the ground is still high, increasing the noise power to counteract it leads to excessive noise injection, reducing transmission rate and communication efficiency. Therefore, the key generation scheme designed in this invention is more secure than the artificial noise scheme.

[0276] Table 1

[0277]

[0278] Finally, the randomness of the physical layer key is evaluated using the NIST standard. When the key passes the randomness test, it is considered to have passed the randomness test. As shown in Table 1, the NIST randomness test results indicate that the key generated by this invention has passed the NIST randomness test.

[0279] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for secure physical layer communication for unmanned aerial vehicles based on orthogonal time-frequency space, characterized in that, Includes the following steps: Step 1: Establish an UAV air-to-ground communication system consisting of a multi-antenna UAV base station, a single-antenna authorized ground user authorized to receive information, and an unauthorized ground user passively eavesdropping. Then, perform time-delay Doppler angle dimension channel modeling on the channel of the UAV air-to-ground communication system to obtain the time-delay Doppler domain channel matrix. Step 2: Using a three-dimensional structured orthogonal matching pursuit algorithm, based on the communication channel between a multi-antenna UAV base station and a single-antenna legitimate user, the index set of the dominant path of the communication channel in the delay dimension, Doppler dimension, and angle dimension is obtained. The channel estimation vector of all dominant paths is determined by the least squares method. Step 3: Adaptively weight the path gain strength of the channel estimation vector of the dominant path to obtain the channel characteristic value of the communication channel, which is used to improve the reciprocity of the communication channel between the multi-antenna UAV base station and the single-antenna legitimate user, and ensure that the two communicating parties can generate consistent physical layer keys. Step 4: Perform double-bit uniform quantization on the channel feature values ​​of the communication channel to generate a channel feature value sequence, and then encode the channel feature value sequence to obtain the initialized bit sequence. Step 5: Perform Winnow key negotiation based on dynamic grouping on the initialized bit sequence. This results in a consistent physical layer key for encrypting the transmitted information between the multi-antenna UAV base station and the single-antenna legitimate user. This improves the bit error rate of unauthorized users, prevents them from eavesdropping, and achieves secure communication using the physical layer key.

2. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in claim 1, characterized in that, Step one specifically includes the following steps: Step 11: First, model the channel of the UAV air-to-ground communication system. Since the downlink channel... The channel estimation method is the same for all single-antenna legitimate users, so channel estimation is only performed for any single-antenna legitimate user. The specific steps are as follows: Definition and the first of multi-antenna UAV base stations The time-domain correlation time-varying channel of each antenna for: ， In the formula, For time-domain discretization indexes, For symbol period, Subcarrier spacing; The sampling interval is... The number of sampling points within the symbol period. An index for time-delay domain discretization; Number of main paths in time-varying channels Each main path contains Subpath, the first The Doppler frequency shift of each subpath is , For the flight speed of multi-antenna drone base stations, For carrier frequency, The speed of light; the channel gain is expressed as , No. The latency of all sub-paths under the main path is 1. Antenna quantity index , This represents the total number of antennas in a multi-antenna drone base station. No. Antenna launch angle of each sub-path for Considering a linear uniform antenna array, and The relevant spatial angle is , The distance between the antennas. The wavelength of the subcarrier; This is the response of the pulse shaping filter; Then we get the first Equivalent Channel Response Matrix in Delay-Doppler Domain on a Single Antenna for: ， In the formula, variables , For the first Doppler frequency shift of each sub-path The number of discrete symbols in the time delay domain. Symbolic index for the Doppler field, Symbolic index for time-delay discrete domain right Take the mold. For the first The latency of all sub-paths under the main path. For the first The antenna spatial angle of each subpath, with respect to variables function for: ， In the formula, The number of symbols in the Doppler field. For summation index variables; Step 12: In order to perform sparse representation of the channel of the UAV air-to-ground communication system, the equivalent channel response matrix in the time-delay Doppler domain is calculated along the spatial domain. Performing the inverse discrete Fourier transform yields the time-delay Doppler angle-dimensional channel. : ， In the formula, For angle index, This represents the total number of antennas in a multi-antenna drone base station. Step 13: In the time-delay Doppler angle dimension channel After modeling, the received signal of a legitimate user with a single antenna With time-delay Doppler domain channel matrix They are represented as follows: ， ， In the formula, Pilot matrix for drone base station transmission, for The identity matrix, The number of drone base station antennas. It is a sparse basis matrix. It is additive white Gaussian noise. It represents the Kronecker product.

3. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in claim 1, characterized in that, Step two specifically includes the following steps: Step 21: Set the loop count index for the 3D structured orthogonal matching pursuit algorithm. Dominant path index set of the time-delay Doppler domain channel matrix The channel estimation vector of the initial time-delay Doppler domain channel matrix Initial residual for: ， In the formula, It is the received signal of a legitimate user with a single antenna. It is the time-delay Doppler domain channel matrix; Step 22: Based on the three-dimensional structured orthogonal matching pursuit algorithm, iterate cyclically according to the number of paths to be estimated, and then calculate the correlation vector of the time-delay Doppler domain channel matrix. for: ， In the formula, The conjugate transpose of the time-delay Doppler domain channel matrix. For the first The residual of the next iteration; Furthermore, the relevant vectors Rearranged into a three-dimensional tensor with delay, Doppler, and antenna dimensions. for: ， In the formula, The length of the protection interval in the delay dimension. The length of the protection interval in the Doppler dimension. The number of antennas for a multi-antenna drone base station; Step 23: Calculate the three-dimensional tensor Expanding along the delay dimension yields the expanded matrix. for: ， Calculate the row vectors of the expanded matrix. Norm, to obtain the correlation vector of the delay dimension Select The delay dimension index corresponding to the maximum value in the middle ,Then As the first The support set for the next iteration in the delay dimension; Step 24: Fixed Delay Dimension Index The first slice matrix of the expanded matrix is: ， Then, the row vectors of the first slice matrix along the Doppler dimension are calculated. Norm, to obtain the Doppler correlation vector By selecting the maximum value in the Doppler correlation vector, the non-zero block in the Doppler dimension is determined, and thus the first... The support set of the next iteration in the Doppler dimension ; Step 25: Based on the first slice matrix and the support set in the Doppler dimension The second slice matrix is ​​obtained as follows: ， Calculate the column vectors of the second slice matrix along the angular dimension. Norm, to obtain the angular dimension correlation vector ; Furthermore, by utilizing the lifting transformation method, the angle-dimensional correlation vector is... Transformation and processing are performed to determine the first... The support set of the next iteration in the angular dimension ; Step 26: Support and update the delay dimension, Doppler dimension, and angle dimension determined in steps 23 to 25 to obtain the index set. for: ， Therefore, the channel estimation vector is obtained by using the least squares method. for: ， Next, update the residual to This is used to eliminate the influence of the estimated dominant paths on the residuals. Steps 23 to 26 are repeated until all dominant paths have been detected and estimated. Finally, the channel estimation vector of the time-delay Doppler domain channel matrix is ​​output. for: ， In the formula, The number of dominant paths.

4. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in claim 1, characterized in that, Step three specifically includes the following steps: Step 31, the channel estimation vector is denoted as: ， In the formula, The number of dominant paths; Step 32: Then, based on the magnitude of the dominant path gain strength, sort the dominant path gain strengths from smallest to largest to obtain: ； In the formula, This is an estimate of the dominant path with the smallest gain intensity after sorting. For the sorted number Gain strength estimates for each dominant path, This is the estimated value of the dominant path with the largest gain after sorting; Step 33: Adaptively weight the dominant path gain strength from 0 to 1 after sorting, i.e.: ， In the formula, For the first Each weight is used to adaptively weight the gain strength of the ranked dominant paths, and its value ranges from 0 to 1. The path corresponding to the minimum gain. The intermediate weight corresponding to the path with the maximum gain Increasing linearly; Step 34: Restore to the original order, perform sliding windowing processing, and obtain the channel characteristic values ​​of the communication channel. : ， In the formula, For window size, The number of channel eigenvalues. This represents the overlap size between two adjacent windows. For the first The size of the adaptive weighted weights, Indicates the first An adaptively weighted estimate of the dominant path gain strength.

5. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in claim 1, characterized in that, Step four specifically includes the following steps: Step 41: The channel feature value sequences for multi-antenna UAV base stations and single-antenna legitimate users are as follows: and The sample size is ; Then, the multi-antenna UAV base station calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ; A single-antenna legitimate user calculates the sequence to be quantized based on its own channel feature value sequence. The sample interval is ; Based on the number of quantization bits Calculate the number of quantized regions Next, the sample intervals are divided into equal intervals as follows: ， ， In the formula, and The classification intervals are for multi-antenna drone base stations and single-antenna legitimate users, respectively. Step 42: Multi-antenna drone base stations and single-antenna legitimate users are classified by interval. and Calculate the quantization threshold, and then use the quantization threshold to quantize the sequence. and In Each sample was classified separately. Step 43: The multi-antenna UAV base station and the single-antenna legitimate user encode the classified samples using Gray code to obtain the initialized bit sequence. and .

6. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in claim 1, characterized in that, Step five specifically includes the following steps: Step 51: The multi-antenna drone base station and the single-antenna legitimate user will initialize the bit sequence. and Calculate the initial bit sequence by randomly permuting it according to the same rules. and Bit inconsistency rate And set a bit inconsistency rate threshold. ; like Then the initialized bit sequence after random permutation and according to Length grouping, Indicates the initial bit sequence Length; like Then the initialized bit sequence after random permutation and according to Length groups are formed, parity bits for each group are calculated, and the groups are exchanged on a common channel. Step 52: Compare the parity check bits exchanged between the multi-antenna UAV base station and the single-antenna legitimate user. If any of the parity check bits are different, perform error correction only on the packets with different parity check bits, discard the first bit of the packet, and use the remaining bits to form a Hamming code group. and Perform local error correction; Step 53: Calculate the Hamming code group and The accompanying , They are respectively: ， ， In the formula, For the verification matrix, This is the conjugate transpose of the parity check matrix; Step 54: The multi-antenna UAV base station transmits the associated signal through a public channel. Send to a single-antenna legitimate user, who then performs a bitwise XOR calculation. ,in The sign is calculated by XORing, thereby obtaining the position of the erroneous bit and realizing a round of dynamic negotiation and error correction; Step 55: Repeat steps 51 to 54 until the initial keys of the multi-antenna drone base station and the single-antenna legitimate user are completely identical, thus obtaining the physical layer key. .

7. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in any one of claims 1-6, characterized in that, To improve the randomness and security of the physical layer key, the method further includes a step of performing SHA-2 hash function encryption on the negotiated physical layer key.

8. The UAV physical layer key-secure communication method based on orthogonal time-frequency space as described in claim 7, characterized in that, The specific steps for enhancing the security of the negotiated physical layer key using the SHA-2 hash function are as follows: Step 61: Based on the SHA-2 algorithm, the initial physical layer key is consistent between the multi-antenna drone base station and the single-antenna legitimate user after key negotiation. Fill in the data, and then set an initial hash value, which consists of eight 32-bit fixed constants; Step 62: The multi-antenna drone base station and the single-antenna legitimate user further divide the message into message blocks, and then perform expansion and round-iteration operations on each message block. In each round, logical functions and bit operations are used to gradually integrate the message block information into the intermediate hash value. Step 63: After processing all message blocks, add the initial hash value and the intermediate results during processing to each hash value in 32-bit increments to produce a final 256-bit hash value, which is the physical layer key enhanced with the SHA-2 hash function. .

Images