A secure communication system for drones

Abstract

The present invention discloses a confidential communication system for unmanned aerial vehicles (UAVs), and relates to the technical field of confidential communication for UAVs. The present invention includes key generation and distributed storage, dynamic key update, key destruction and tracing, multi-band perception and interference identification, and a joint control unit. Dynamic keys are generated by a chaotic algorithm, combined with blockchain shard storage and GPS binding encryption to ensure key security; keys are updated according to risk levels, and shards are transmitted using quantum-resistant channels; key shards are invalidated after the communication is completed to achieve anomaly detection and tracing; multi-band spectrum is scanned in real time to identify interference types and optimize models; frequency bands are dynamically switched, power is adjusted, and anti-interception capabilities are enhanced based on interference intensity and energy efficiency requirements. The present invention effectively solves the problems of insufficient security, weak anti-interference capabilities, and imperfect key management mechanisms in existing UAV communication systems, and significantly improves the security, anti-interference performance, and energy efficiency of UAV communications.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) secure communication, and in particular to a UAV secure communication system. Background Art

[0002] With the rapid development of drone technology, its application in both civil and commercial fields is becoming increasingly widespread. However, the security and confidentiality of drone communications face severe challenges. Traditional drone communication systems often use a single frequency band for transmission, which is susceptible to interference and interception. This makes it difficult to ensure the stability and security of communication links, especially in complex electromagnetic environments or hostile environments. Furthermore, existing communication systems have numerous deficiencies in key management. For example, inflexible key distribution, storage, and update mechanisms make them susceptible to cracking or leakage, leading to the theft or tampering of communication content.

[0003] In modern communication environments, drones must contend with a variety of interference and attack vectors, including narrowband interference, broadband noise interference, and active attacks from malicious nodes. While existing technologies offer some anti-interference methods, most can only address specific types of interference and lack the ability to comprehensively identify and dynamically address multiple types of interference. Furthermore, traditional communication systems also struggle with energy efficiency optimization, unable to dynamically adjust power and frequency bands based on the real-time communication environment and mission requirements. This results in low communication efficiency and high energy consumption. Furthermore, existing systems lack effective risk assessment and dynamic adjustment mechanisms during key management, making it impossible to promptly update keys based on changes in the network environment, further increasing the risk of communication compromise.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] The purpose of the present invention is to solve the problems of insufficient security, weak anti-interference ability and imperfect key management mechanism in existing drone communication systems, and to propose a confidential communication system for drones.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A secure communication system for an unmanned aerial vehicle, comprising:

[0008] Key generation and distributed storage unit generates dynamic keys through chaos algorithm, combines blockchain sharding storage with GPS binding encryption to achieve secure key distribution and attack-resistant storage;

[0009] Dynamic key update unit, which triggers key updates based on risk level, generates new keys using PBFT consensus, and transmits shards over quantum-resistant channels;

[0010] The key destruction and tracing unit is used to invalidate key shards after the communication ends, and combines machine learning with blockchain logs to achieve anomaly detection and tracing;

[0011] Multi-band sensing and interference identification unit, which scans multi-band spectrum in real time, classifies interference types using a lightweight CNN, and optimizes the model through online learning;

[0012] The joint control unit is used to dynamically switch frequency bands, adjust power according to interference intensity and energy efficiency requirements, and enable beamforming to enhance anti-interception capabilities.

[0013] Furthermore, the key generation and distributed storage unit execution process is as follows:

[0014] A random key seed is generated using a chaotic sequence algorithm, and the initial key is generated by combining the drone's unique identification code. ;

[0015] Through the hash algorithm Encrypt and generate the final key ;

[0016] Using Shamir's secret sharing algorithm Split into n shards, and store the shards on multiple nodes of the blockchain network; and use dynamic sharding strategies to adaptively adjust the number of shards;

[0017] Use Shamir's secret sharing algorithm to When sharding, double encryption is introduced:

[0018] The first layer of encryption: for each shard ; Second layer of encryption: Bind to the current position coordinates of the drone to generate the final fragment ;

[0019] Allocate shard storage locations by risk level: Public chain node storage ; Alliance chain node storage ; Shard storage metadata including shard ID and storage location is recorded and encrypted by the smart contract;

[0020] When the drone and the ground station need to establish communication, they submit to the smart contract: digital signatures of both parties; the current GPS coordinates are used to decrypt the shard; the smart contract verifies: the validity of the signature; and whether the deviation between the GPS coordinates and the coordinates stored in the shard is within the allowable range, which is ≤10km.

[0021] The smart contract retrieves valid shards from the public chain and the consortium chain based on metadata; if the number of available shards is ≥ k, the reorganization process begins; otherwise, a key update is triggered; for each shard Execution: Restore using k shards through the Shamir algorithm ;

[0022] Periodic update: Force an update of the sharding rules every 24 hours, even if R has not changed; Event-driven update: Immediately trigger an update when the following events are detected: The number of failed shard calls > 3 times; The drone position has moved more than 100 km; The credibility of the blockchain node has decreased by more than 20%; Generate new sharding rules; Mark the old shards as invalid and re-store the new shards according to the rules.

[0023] Furthermore, the specific process of the dynamic sharding strategy is as follows:

[0024] Real-time collect the following parameters through the smart contract to evaluate the network risk level R:

[0025] Node credibility: Based on historical behavior scores, including shard call compliance and response latency;

[0026] Attack frequency: The number of abnormal access requests detected per unit time;

[0027] Cross-chain communication latency: The data synchronization time between the public chain and the consortium chain;

[0028] Automatically adjust the sharding parameters according to the risk level R, and the process is as follows:

[0029] When the risk level R ≤ 0.3, it is judged as a low risk, then the number of shards n = 3, and the recombination threshold k = 2; 60% of the shards are stored in the public chain and 40% are stored in the consortium chain for the shard storage location;

[0030] When the risk level 0.3 < R ≤ 0.7, it is judged as a medium risk, then n = 5, k = 3, and 50% of the shards are stored in each of the public chain and the consortium chain;

[0031] When the risk level R > 0.7, it is judged as a high risk, n = 7, k = 4, and 30% of the shards are stored in the public chain and 70% are stored in the consortium chain;

[0032] Sharding rule formula: ; where n represents the number of shards; k represents the minimum number of shards required to recombine the key; R represents the network risk level.

[0033] Furthermore, the specific operation steps of the dynamic key update unit are as follows:

[0034] When the drone and the ground station initiate a communication request, submit the following information to the blockchain network:

[0035] Digital signature: A two-party signature based on the ECDSA algorithm to verify the legitimacy of the identity; Real-time risk level R1: Obtain the current network risk level R1 from the blockchain network and analyze it based on the proportion of malicious nodes and attack frequency;

[0036] Smart contracts perform two-factor authentication, including identity verification: checking signature validity to ensure both parties are legitimate devices; risk level threshold judgment: when R1 ≥ 0.5, a mandatory key update is triggered; when R1 < 0.5, a periodic update is triggered only when the communication interval exceeds the set time;

[0037] At the same time, the number of PBFT consensus nodes is adjusted based on the real-time risk level R1. When R1 is less than 0.3, 3 nodes participate in the consensus; when R1 is greater than or equal to 0.3, 7 nodes participate in the consensus. The consensus group nodes must meet the historical credibility score of ≥80% and be physically dispersed.

[0038] The consensus group generates a new key using the PBFT algorithm ;Specific process:

[0039] Proposal stage: The master node generates a random number seed and broadcasts it to other nodes;

[0040] Verification phase: Nodes verify the legitimacy of the seed and generate it after voting. ;

[0041] Key sharding: adopt dynamic sharding strategy and adjust the number of shards n and threshold k according to R;

[0042] The new key shard is transmitted to the drone and ground station via a quantum secure channel:

[0043] Sharded data encryption: Use NTRU quantum-resistant encryption algorithm to protect the transmission process;

[0044] Dynamic transmission path selection: Select the optimal blockchain node path based on real-time network latency.

[0045] Furthermore, the specific operation steps of the key destruction and tracing unit are as follows:

[0046] After completing the communication task, the drone sends an encrypted termination instruction to the blockchain network, which includes: the communication session ID (hash value), the current key version number, and the dual digital signatures of the drone and the ground station;

[0047] The smart contract performs the following operations:

[0048] Related shard retrieval: locate all related key shards based on the session ID and key version number;

[0049] Shard status update: mark the shard as invalid and write it to the blockchain;

[0050] Storage isolation: Failed shards remain on-chain but are moved to a read-only storage area, prohibiting any call operations;

[0051] Smart contracts monitor the following metrics in real time:

[0052] Shard call frequency: the number of shard calls per unit time;

[0053] Access source IP: detects access to nodes from unusual geographical locations;

[0054] Node behavior patterns: Machine learning models are used to identify abnormal behaviors, including abnormal shard call timing.

[0055] Different abnormal events are classified and responses are triggered according to different threat levels, including:

[0056] Low-level exceptions: logs are recorded but no alarm is issued; medium-level exceptions: shard call permissions are frozen and the administrator is notified; high-level exceptions: identified as attacks, triggering emergency key destruction;

[0057] For abnormal node location, combined with the node registration information stored in the blockchain, the physical device corresponding to the abnormal IP is located, the transaction hash, digital signature, and timestamp of the abnormal operation are extracted, and an unalterable forensic report is generated;

[0058] At the same time, the smart contract automatically executes, including adding abnormal nodes to the blacklist, prohibiting them from participating in subsequent consensus or shard storage, and triggering the key update unit to immediately generate new keys to replace the attacked keys, and push alarm information to the security center.

[0059] Furthermore, the specific operation steps of the multi-band sensing and interference identification unit are as follows:

[0060] The drone is equipped with a wide-band spectrum scanning chip, which covers L-band 1-2GHz, C-band 4-8GHz, and Ku-band 12-18GHz, and scans the spectrum every 10ms;

[0061] Collect signal strength and noise power parameters of each frequency band;

[0062] Apply wavelet transform to the original spectrum data to filter out the environmental background noise; generate the spectrum feature matrix, including: peak frequency, bandwidth, instantaneous power change rate; frequency domain entropy;

[0063] Interference classification is performed using a pre-trained lightweight CNN model. The specific structure of the pre-trained lightweight CNN model is an input layer [128x128 spectrum map], 3 layers of convolution + 2 layers of full connection: Input normalization: normalize the spectrum data to the range of [-1, 1];

[0064] Interference type determination: Narrowband interference: output probability > 0.8 and meets the following conditions: main lobe width < 5MHz, peak power > -70dBm; instantaneous power change rate > 10dB / ms;

[0065] Broadband noise interference: output probability > 0.7 and meets the following requirements: spectrum flatness > 0.9, standard deviation < 2dB; frequency domain entropy > 6.0;

[0066] Interference intensity quantification: divided into levels according to the signal-to-noise power ratio: low intensity is SNR < 10dB, medium intensity is 10dB ≤ SNR ≤ 20dB, and high intensity is SNR > 20dB;

[0067] Incremental training: If an unknown interference pattern is detected, that is, the CNN confidence is less than 0.6, the following process is triggered:

[0068] Collect new spectrum data and annotate it, either manually or through a rule-based engine. Fine-tune the model and update the convolution kernel weights on the edge computing unit. Synchronize the new model parameters to the blockchain network for download and verification by other drone nodes.

[0069] The interference result output and response process are as follows:

[0070] Generate interference maps and construct spatiotemporal interference heat maps, including: interference type, intensity, and duration for each frequency band; and the location of the interference source. The heat maps are uploaded to the ground command center via an encrypted channel.

[0071] Real-time anti-interference strategy triggering, generating control instructions based on recognition results:

[0072] Narrowband interference: triggers the dynamic frequency band switching unit to hop to the interference-free sub-band;

[0073] Broadband interference: triggers the multi-band cooperative transmission unit to start redundant coding;

[0074] High-intensity jamming: Activate the decoy signal transmission module to attract enemy jamming resources;

[0075] The spectrum scanning chip adopts electromagnetic shielding packaging to prevent external interference injection; it has a built-in self-test circuit to regularly check the ADC sampling accuracy; if the scanning data is abnormal for three consecutive times, it switches to the backup spectrum database.

[0076] Furthermore, the specific operation steps of the joint control unit are as follows:

[0077] Synchronously acquire spectrum status, channel quality, and environmental parameters, and generate a spectrum-power joint decision matrix;

[0078] The frequency band selection and transmission mode switching process is as follows:

[0079] Dynamic election of primary frequency band: Frequency band scoring model: score for each frequency band Where S is the frequency band score; SNR is the signal-to-noise ratio; the frequency band with the highest score is selected as the primary channel. If its interference strength exceeds the threshold of -90dBm, the backup frequency bands are activated in sequence, i.e. C band → L band;

[0080] Multi-band coordinated transmission activation: Trigger conditions are: interference intensity of all bands > -85dBm and BER > 1e-4;

[0081] Data segmentation and encoding: The data packet is divided into three sub-streams, with one sub-stream each carried by the L-band, C-band, and Ku-band. Reed-Solomon redundancy coding is added to each sub-stream. Parallel transmission: The three sub-streams are transmitted synchronously via software-defined radio.

[0082] The adaptive power-beam coordinated control process is as follows:

[0083] Power dynamic calculation: Free space loss compensation: According to the formula Calculate theoretical loss; where L represents free space loss; d represents communication distance; λ represents signal wavelength; dynamically adjust transmit power based on loss value L , so that the received power Maintain above -80dBm;

[0084] When the interference intensity is greater than -80dBm and the stealth requirement level is ≥2, the target direction angle is obtained through the phased array antenna; the beam pointing angle θ is calculated, and the array element phase is adjusted to align the main lobe with the receiving end; a null is generated in the interference direction to suppress the reception of the interference signal.

[0085] Compared with the prior art, the present invention has the following beneficial effects:

[0086] (1) The present invention significantly improves the security of UAV communications through innovative key management and storage mechanisms. The key generation and distributed storage unit uses a chaotic algorithm to generate dynamic keys, and combines blockchain sharding storage with GPS-bound encryption to achieve secure key distribution and attack-resistant storage. At the same time, the dynamic key update unit triggers key updates based on risk levels and generates new keys through PBFT consensus, ensuring that the keys always maintain a high level of security during the communication process. In addition, the key destruction and tracing unit can invalidate the key shards after the communication ends, and combines machine learning with blockchain logs to achieve anomaly detection and traceability, further ensuring the confidentiality and security of communications.

[0087] (2) In the present invention, the multi-band sensing and interference identification unit can scan the multi-band spectrum in real time, and use lightweight CNN to classify interference types. By optimizing the model through online learning, the accuracy and real-time performance of interference identification are effectively improved; the joint control unit dynamically switches the frequency band and adjusts the power according to the interference intensity and energy efficiency requirements, and enables beamforming to enhance the anti-interception capability, which significantly improves the anti-interference performance and energy efficiency of UAV communication; the comprehensive anti-interference and energy efficiency optimization strategy enables the UAV communication system to maintain stable and efficient communication capabilities in complex electromagnetic environments;

[0088] (3) This invention combines blockchain technology, quantum-resistant encryption algorithms, dynamic key management, multi-band sensing and interference identification and other advanced technologies to build a highly secure, flexible and efficient drone confidential communication system; it can not only effectively deal with various threats in the modern communication environment, but also has good adaptability and scalability, and can meet the high security requirements of drone communications in both civil and commercial fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0090] Figure 1 This is the overall system block diagram of the present invention. DETAILED DESCRIPTION

[0091] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0092] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0093] It should also be understood that the terminology used in this disclosure is for the purpose of describing specific embodiments only and is not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should be further understood that the term "and / or" as used in this disclosure and the claims refers to any and all possible combinations of one or more of the associated listed items, including and including these combinations.

[0094] As shown Figure 1 A secure communication system for an unmanned aerial vehicle (UAV) includes a key generation and distributed storage unit, a dynamic key update unit, a key destruction and traceability unit, a multi-band sensing and interference recognition unit, and a joint regulation unit;

[0095] The key generation and distributed storage unit generates dynamic keys through a chaotic algorithm, combines blockchain sharding storage and GPS binding encryption to achieve secure key distribution and anti-attack storage;

[0096] Generate a random key seed (such as the Lorenz chaos model) using a chaotic sequence algorithm, and generate an initial key in combination with the unique identification code of the UAV (such as the IMEI code) ; Through a hash algorithm (such as SHA-256) for to encrypt and generate the final key ; Use the Shamir secret sharing algorithm to be divided into n shards (for example, n = 5), and the shards are stored in multiple nodes of the blockchain network; and use a dynamic sharding strategy to adaptively adjust the number of shards. The specific process of the dynamic sharding strategy is as follows:

[0097] Real-time collect the following parameters through a smart contract to evaluate the network risk level R: Node credibility: Based on historical behavior scores (such as shard call compliance, response latency); Attack frequency: The number of abnormal access requests detected per unit time; Cross-chain communication latency: The data synchronization time between the public chain and the consortium chain (threshold: ≤200ms); Automatically adjust the sharding parameters according to the risk level R:

[0098] Low risk (R ≤ 0.3): The number of shards n = 3, the recombination threshold k = 2; Sharding storage location: 60% of the shards are stored in the public chain, 40% are stored in the consortium chain; Medium risk (0.3 < R ≤ 0.7): n = 5, k = 3, and 50% of the shards are stored in each of the public chain and the consortium chain; High risk (R > 0.7): n = 7, k = 4, 30% of the shards are stored in the public chain, 70% are stored in the consortium chain; Sharding rule formula: ; Where n represents the number of shards; k represents the minimum number of shards required to recombine the key; R represents the network risk level.

[0099] When using the Shamir secret sharing algorithm to shard the key, introduce double encryption:

[0100] The first layer of encryption: For each shard ; The second layer of encryption: Bind with the current position coordinates of the UAV (GPS data) to generate the final shard ; Allocate the sharding storage location according to the risk level: The public chain nodes store (Example: IPFS network); Alliance chain node storage (Hyperledger Fabric private channel); shard storage metadata (shard ID, storage location) is recorded and encrypted by the smart contract; when the drone and the ground station need to establish communication, they submit to the smart contract: digital signatures of both parties (based on the ECDSA algorithm); current GPS coordinates (used to decrypt the shard); the smart contract verifies: signature validity; whether the deviation between the GPS coordinates and the coordinates when the shard is stored is within the allowable range (≤10km);

[0101] The smart contract retrieves valid shards (not marked as invalid) from the public chain and the consortium chain based on metadata; if the number of available shards is ≥ k, it enters the reorganization process; otherwise, it triggers the key update (see the dynamic key update unit); for each shard Execution: Restore using Shamir's algorithm using k shards ; Periodic update: sharding rules are forced to be updated every 24 hours (even if R has not changed); event-driven update: updates are triggered immediately when the following events are detected: number of sharding call failures > 3; drone location moves more than 100km; blockchain node credibility drops by more than 20%; new sharding rules are generated (adjust n, k and storage ratio); old shards are marked as invalid, and new shards are re-stored according to the rules.

[0102] The dynamic key update unit triggers key updates based on risk levels, generates new keys using PBFT consensus, and transmits shards over quantum-resistant channels;

[0103] When the drone initiates a communication request with the ground station, it submits the following information to the blockchain network: Digital signature: a signature of both parties based on the ECDSA algorithm to verify the legitimacy of the identity; Real-time risk level R1: obtaining the current network risk level (such as the proportion of malicious nodes and attack frequency) from the blockchain network; Smart contracts perform two-factor verification, including identity verification: checking the validity of the signature to ensure that both parties are legitimate devices; Risk level threshold judgment: When R1 ≥ 0.5 (medium-high risk), a mandatory key update is triggered; when R1 < 0.5, a periodic update is triggered only when the communication interval exceeds the set time;

[0104] At the same time, the number of PBFT consensus nodes is adjusted according to the real-time risk level R1. When R1 is less than 0.3, 3 nodes participate in the consensus; when R1 is greater than or equal to 0.3, 7 nodes participate in the consensus. The consensus group nodes must meet the following requirements: historical credibility score ≥ 80%; physical location is dispersed (to avoid physical attacks on single-region nodes); the consensus group generates new keys through the PBFT algorithm. Specific process: Proposal stage: The master node generates a random number seed and broadcasts it to other nodes; Verification stage: The node verifies the legitimacy of the seed and generates a random number after voting. ;Key sharding: Use dynamic sharding strategy (see key generation unit) and adjust the number of shards n and threshold k according to R;

[0105] The new key shards are transmitted to the drone and ground station via a quantum secure channel (QKD): shard data encryption: the NTRU quantum-resistant encryption algorithm is used to protect the transmission process; dynamic transmission path selection: the optimal blockchain node path is selected based on real-time network latency.

[0106] The key destruction and tracing unit is used to invalidate key shards after the communication ends, and combines machine learning with blockchain logs to achieve anomaly detection and tracing;

[0107] After completing the communication task, the drone sends an encrypted termination instruction to the blockchain network, which includes: the communication session ID (hash value), the current key version number, and the dual digital signature (ECDSA) of the drone and the ground station. The smart contract performs the following operations:

[0108] Related shard retrieval: locate all related key shards (including shards stored in public and consortium chains) based on the session ID and key version number; shard status update: mark the shard as "invalid" and write it to the blockchain (record the expiration timestamp); storage isolation: the invalid shard remains on the chain, but is moved to a read-only storage area, prohibiting any call operations;

[0109] Smart contracts monitor the following metrics in real time:

[0110] Shard call frequency: The number of shard calls per unit time (threshold: ≤3 times / minute); Access source IP: Detects node access from unusual geographical locations (such as a sudden increase in overseas IP addresses); Node behavior pattern: Identifies abnormal behavior (abnormal shard call timing) through machine learning models; Different abnormal events are classified and responses are triggered based on different threat levels, including:

[0111] Low-level anomalies (such as a single frequency limit violation): logs are recorded, but no alarm is issued for the time being; intermediate anomalies (such as IP address anomalies + frequency limit violation): shard call permissions are frozen and the administrator is notified; high-level anomalies (such as shard reorganization failure accompanied by high-frequency access): it is determined to be an attack and triggers emergency key destruction;

[0112] To locate abnormal nodes, combined with the node registration information stored in the blockchain, the physical device corresponding to the abnormal IP is located, the transaction hash, digital signature, and timestamp of the abnormal operation are extracted, and an unalterable forensic report is generated; at the same time, the smart contract automatically executes: the abnormal node is added to the blacklist, prohibiting it from participating in subsequent consensus or shard storage; the key update unit is triggered to immediately generate a new key to replace the attacked key; and an alarm information (including a summary of the forensic report) is pushed to the security center.

[0113] The multi-band sensing and interference identification unit is used to scan the multi-band spectrum in real time, classify interference types using a lightweight CNN, and optimize the model through online learning;

[0114] The drone is equipped with a wide-band spectrum scanning chip (covering L-band 1-2 GHz, C-band 4-8 GHz, and Ku-band 12-18 GHz), which scans the spectrum every 10 ms. It collects signal strength and noise power parameters for each frequency band. It applies a wavelet transform (Daubechies4 wavelet basis) to the raw spectrum data to filter out ambient background noise. It generates a spectrum feature matrix, including peak frequency, bandwidth, instantaneous power change rate, and frequency domain entropy (a measure of spectrum complexity). It uses a pre-trained lightweight CNN model (structure: input layer [128x128 spectrum graph], 3 convolution layers + 2 fully connected layers) to perform interference classification: Input normalization: Normalize the spectrum data to the range [-1, 1]. Interference type determination:

[0115] Narrowband interference: Output probability > 0.8 and meet the following conditions: main lobe width < 5 MHz, peak power > -70 dBm; instantaneous power change rate > 10 dB / ms (consistent with radar pulse characteristics); Broadband noise interference: Output probability > 0.7 and meet the following conditions: Spectral flatness > 0.9 (standard deviation < 2 dB); Frequency domain entropy > 6.0 (high randomness); Interference intensity quantification: Based on the signal-to-noise power ratio (SNR), it is divided into levels: low intensity (SNR < 10 dB), medium intensity (10 dB ≤ SNR ≤ 20 dB), and high intensity (SNR > 20 dB).

[0116] Incremental training: If an unknown interference pattern is detected (CNN confidence < 0.6), the following process is triggered:

[0117] Collect new spectrum data and annotate it (through manual review or rule-based engine-assisted annotation); perform model fine-tuning on the edge computing unit (such as the drone-side FPGA) to update the convolution kernel weights; synchronize the new model parameters to the blockchain network for download and verification by other drone nodes;

[0118] Interference result output and response:

[0119] Generate interference maps and construct spatiotemporal interference heat maps, including: interference type, intensity, and duration in each frequency band; possible location of interference sources (through multi-UAV collaborative spectrum direction finding); heat maps are uploaded to the ground command center via encrypted channels; real-time anti-interference strategy triggering generates control instructions based on identification results: Narrowband interference: trigger the dynamic frequency band switching unit to hop to an interference-free sub-band; Wideband interference: trigger the multi-band collaborative transmission unit to initiate redundant coding (such as Turbo code); High-intensity interference: activate the decoy signal transmission module to attract enemy interference resources;

[0120] The spectrum scanning chip uses electromagnetic shielding packaging to prevent external interference injection; a built-in self-test circuit regularly verifies the ADC sampling accuracy (error <±0.5dB); if the scanning data is abnormal for three consecutive times (such as full-band saturation), it switches to the backup spectrum database (which stores historical clean spectrum templates).

[0121] The joint control unit is used to dynamically switch frequency bands, adjust power according to interference intensity and energy efficiency requirements, and enable beamforming to enhance anti-interception capabilities;

[0122] Synchronously obtain the following data:

[0123] Spectrum status: interference type (narrowband / wideband) and intensity level (low / medium / high) from the multi-band sensing unit; channel quality: bit error rate (BER) and signal-to-noise ratio (SNR) reported by the receiver; environmental parameters: communication distance d (via GPS or laser ranging), current band center frequency f;

[0124] Generate spectrum-power joint decision matrix, weight distribution table 1:

[0125]

[0126] Table 1

[0127] The frequency band selection and transmission mode switching process is as follows:

[0128] Dynamic election of primary frequency band: Frequency band scoring model: score for each frequency band Where S is the frequency band score; SNR is the signal-to-noise ratio; the frequency band with the highest score is selected as the primary channel (for example, Ku band). If its interference strength exceeds a threshold (-90dBm), the backup frequency bands (C band → L band) are activated in sequence; Multi-band coordinated transmission activation: The trigger condition is: the interference strength of all frequency bands is greater than -85dBm and the BER is greater than 1e-4; Data segmentation and encoding: The data packet is divided into three sub-streams (L / C / Ku bands each carry one sub-stream); Reed-Solomon (255,223) redundant coding is added to each sub-stream, which can correct up to 16 bytes of errors; Parallel transmission: The three sub-streams are transmitted simultaneously via software-defined radio (SDR);

[0129] Adaptive power-beam coordinated control is as follows:

[0130] Power Dynamics Calculation: Free Space Loss Compensation:

[0131] According to the formula Calculate theoretical loss; where L represents free space loss; d represents communication distance; λ represents signal wavelength; dynamically adjust transmit power based on loss value L , so that the received power Maintain above -80dBm; when the interference intensity is greater than -80dBm and the stealth requirement level is ≥2 (special mission mode), obtain the target direction angle (based on the GPS coordinates of the ground station) through the phased array antenna; calculate the beam pointing angle θ, adjust the array element phase to align the main lobe with the receiving end; generate a null in the interference direction to suppress the reception of the interference signal (depth ≥20dB).

[0132] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A secure communication system for drones, characterized in that: It includes: A key generation and distributed storage unit that generates dynamic keys through a chaotic algorithm and combines blockchain sharding storage with GPS - bound encryption to achieve secure key distribution and anti - attack storage; A dynamic key update unit that triggers key updates based on risk levels, generates new keys using PBFT consensus, and transmits shards through a quantum - resistant channel; A key destruction and traceability unit that invalidates key shards after communication ends and combines machine learning with blockchain logs to achieve anomaly detection and traceability; A multi - band sensing and interference recognition unit that scans multi - band spectra in real - time, classifies interference types using lightweight CNN, and optimizes the model through online learning; A joint regulation unit that dynamically switches frequencies, adjusts power according to interference intensity and energy efficiency requirements, and enables beamforming to enhance anti - interception capabilities; The execution process of the key generation and distributed storage unit is as follows: A random key seed is generated using a chaotic sequence algorithm, and the initial key is generated by combining the drone's unique identification code. ; Through the hash algorithm Encrypt and generate the final key ; Using Shamir's secret sharing algorithm Split into n shards, and store the shards on multiple nodes of the blockchain network; and use dynamic sharding strategies to adaptively adjust the number of shards; Use Shamir's secret sharing algorithm to When sharding, double encryption is introduced: The first layer of encryption: for each shard ; Second layer of encryption: Bind to the current position coordinates of the drone to generate the final fragment ; Allocate shard storage locations by risk level: Public chain node storage ; Alliance chain node storage ; Shard storage metadata including shard ID and storage location is recorded and encrypted by the smart contract; When the drone needs to establish communication with the ground station, it submits to the smart contract: the digital signatures of both parties; the current GPS coordinates for decrypting shards; the smart contract verifies: the validity of the signatures; whether the deviation between the GPS coordinates and the coordinates at the time of shard storage is within the allowable range, and the allowable deviation range is ≤ 10km; The smart contract retrieves valid shards from the public chain and the consortium chain according to the metadata; if the number of available shards ≥ k, it enters the recombination process; Otherwise, a key update is triggered; for each shard Execution: Restore using Shamir's algorithm using k shards ; Periodic update: Force an update of the sharding rules every 24 hours, even if R has not changed; Event - driven update: Immediately trigger an update when the following events are detected: the number of failed shard calls > 3 times; the drone's position moves more than 100km; the credibility of the blockchain node drops by more than 20%; Generate new sharding rules; Mark the old shards as invalid and store the new shards according to the rules; The specific process of the dynamic sharding strategy is as follows: Evaluate the network risk level R by collecting the following parameters in real - time through the smart contract: Node credibility: Scored based on historical behavior, including shard call compliance and response latency; Attack frequency: The number of abnormal access requests detected per unit time; Cross - chain communication latency: The data synchronization time between the public chain and the consortium chain; Automatically adjust sharding parameters according to the risk level R, and the process is as follows: When the risk level R ≤ 0.3, it is judged as a low - risk level, then the number of shards n = 3, and the recombination threshold k = 2; 60% of the shards are stored in the public chain and 40% are stored in the consortium chain for shard storage location selection; When the risk level 0.3 < R ≤ 0.7, it is judged as a medium - risk level, then n = 5, k = 3, and 50% of the shards are stored in each of the public chain and the consortium chain; When the risk level R > 0.7, it is judged as a high - risk level, n = 7, k = 4, and 30% of the shards are stored in the public chain and 70% are stored in the consortium chain; Sharding rule formula: ; where n is the number of shards; k is the minimum number of shards required to reconstruct the key; R is the network risk level.

2. The secure communication system for an unmanned aerial vehicle according to claim 1, characterized in that: The specific operation steps of the dynamic key update unit are as follows: When the drone and the ground station initiate a communication request, they submit the following information to the blockchain network: Digital signature: The signatures of both parties based on the ECDSA algorithm to verify the legitimacy of the identity; Real - time risk level R1: Obtain the current network risk level R1 from the blockchain network and analyze it through the proportion of malicious nodes and attack frequency; Smart contracts perform two-factor authentication, including identity verification: checking signature validity to ensure both parties are legitimate devices; and risk level threshold determination: when R1 ≥ 0.5, a mandatory key update is triggered. When R1 is less than 0.5, periodic updates are triggered only when the communication interval exceeds the set time; At the same time, the number of PBFT consensus nodes is adjusted based on the real-time risk level R1. When R1 is less than 0.3, 3 nodes participate in the consensus; when R1 is greater than or equal to 0.3, 7 nodes participate in the consensus. The consensus group nodes must meet the historical credibility score of ≥80% and be physically dispersed. The consensus group generates a new key using the PBFT algorithm ; Specific process: Proposal stage: The master node generates a random number seed and broadcasts it to other nodes; Verification phase: Nodes verify the legitimacy of the seed and generate it after voting. ; Key sharding: adopt dynamic sharding strategy and adjust the number of shards n and threshold k according to R; The new key shard is transmitted to the drone and ground station via a quantum secure channel: Sharded data encryption: Use NTRU quantum-resistant encryption algorithm to protect the transmission process; Dynamic transmission path selection: Select the optimal blockchain node path based on real-time network latency.

3. The secure communication system for an unmanned aerial vehicle according to claim 1, characterized in that: The specific operation steps of the key destruction and tracing unit are as follows: After completing the communication task, the drone sends an encrypted termination instruction to the blockchain network, which includes: the communication session ID (hash value), the current key version number, and the dual digital signatures of the drone and the ground station; The smart contract performs the following operations: Related shard retrieval: locate all related key shards based on the session ID and key version number; Shard status update: mark the shard as invalid and write it to the blockchain; Storage isolation: Failed shards remain on-chain but are moved to a read-only storage area, prohibiting any call operations; Smart contracts monitor the following metrics in real time: Shard call frequency: the number of shard calls per unit time; Access source IP: detects access to nodes from unusual geographical locations; Node behavior patterns: Machine learning models are used to identify abnormal behaviors, including abnormal shard call timing. Different abnormal events are classified and responses are triggered according to different threat levels, including: Low-level exceptions: logs are recorded but no alarm is issued; medium-level exceptions: shard call permissions are frozen and the administrator is notified; high-level exceptions: identified as attacks, triggering emergency key destruction; For abnormal node location, combined with the node registration information stored in the blockchain, the physical device corresponding to the abnormal IP is located, the transaction hash, digital signature, and timestamp of the abnormal operation are extracted, and an unalterable forensic report is generated; At the same time, the smart contract automatically executes, including adding abnormal nodes to the blacklist, prohibiting them from participating in subsequent consensus or shard storage, and triggering the key update unit to immediately generate new keys to replace the attacked keys, and push alarm information to the security center.

4. The secure communication system for an unmanned aerial vehicle according to claim 1, characterized in that: The specific operation steps of the multi-band sensing and interference identification unit are as follows: The drone is equipped with a wide-band spectrum scanning chip, which covers L-band 1-2GHz, C-band 4-8GHz, and Ku-band 12-18GHz, and scans the spectrum every 10ms; Collect signal strength and noise power parameters of each frequency band; Apply wavelet transform to the original spectrum data to filter out the environmental background noise; generate Spectrum feature matrix, including: peak frequency, bandwidth, instantaneous power change rate; frequency domain entropy; Interference classification is performed using a pre-trained lightweight CNN model. The specific structure of the pre-trained lightweight CNN model is an input layer [128x128 spectrum map], 3 layers of convolution + 2 layers of full connection: Input normalization: normalize the spectrum data to the range of [-1, 1]; Interference type determination: Narrowband interference: output probability > 0.8 and meets the following conditions: main lobe width < 5MHz, peak power > -70dBm; instantaneous power change rate > 10dB / ms; Broadband noise interference: output probability > 0.7 and meets the following requirements: spectrum flatness > 0.9, standard deviation < 2dB; frequency domain entropy > 6.0; Interference intensity quantification: divided into levels according to the signal-to-noise power ratio: low intensity is SNR < 10dB, medium intensity is 10dB ≤ SNR ≤ 20dB, and high intensity is SNR > 20dB; Incremental training: If an unknown interference pattern is detected, that is, the CNN confidence is less than 0.6, the following process is triggered: Collect new spectrum data and annotate it, either manually or through a rule-based engine. Fine-tune the model and update the convolution kernel weights on the edge computing unit. Synchronize the new model parameters to the blockchain network for download and verification by other drone nodes. The interference result output and response process are as follows: Generate interference maps and construct spatiotemporal interference heat maps, including: interference type, intensity, and duration for each frequency band; and the location of the interference source. The heat maps are uploaded to the ground command center via an encrypted channel. Real-time anti-interference strategy triggering, generating control instructions based on recognition results: Narrowband interference: triggers the dynamic frequency band switching unit to hop to the interference-free sub-band; Broadband interference: triggers the multi-band cooperative transmission unit to start redundant coding; High-intensity jamming: Activate the decoy signal transmission module to attract enemy jamming resources; The spectrum scanning chip adopts electromagnetic shielding packaging to prevent external interference injection; it has a built-in self-test circuit to regularly check the ADC sampling accuracy; if the scanning data is abnormal for three consecutive times, it switches to the backup spectrum database.

5. The secure communication system for an unmanned aerial vehicle according to claim 1, characterized in that: The specific operation steps of the joint control unit are as follows: Synchronously acquire spectrum status, channel quality, and environmental parameters, and generate a spectrum-power joint decision matrix; The frequency band selection and transmission mode switching process is as follows: Dynamic election of primary frequency band: Frequency band scoring model: score for each frequency band Where S is the frequency band score; SNR is the signal-to-noise ratio; the frequency band with the highest score is selected as the primary channel. If its interference strength exceeds the threshold of -90dBm, the backup frequency bands are activated in sequence, i.e. C band → L band; Multi-band coordinated transmission activation: Trigger conditions are: interference intensity of all bands > -85dBm and BER > 1e-4; Data segmentation and encoding: The data packet is divided into three sub-streams, with one sub-stream each carried by the L-band, C-band, and Ku-band. Reed-Solomon redundancy coding is added to each sub-stream. Parallel transmission: The three sub-streams are transmitted synchronously via software-defined radio. The adaptive power-beam coordinated control process is as follows: Power dynamic calculation: Free space loss compensation: According to the formula Calculate theoretical loss; where L represents free space loss; d represents communication distance; dynamically adjust transmit power based on loss value L , so that the received power Maintain above -80dBm; When the interference intensity is greater than -80dBm and the stealth requirement level is ≥2, the target direction angle is obtained through the phased array antenna; the beam pointing angle θ is calculated, and the array element phase is adjusted to align the main lobe with the receiving end; a null is generated in the interference direction to suppress the reception of the interference signal.

Images