Distributed information safety protection device for unmanned vehicle

Abstract

The invention discloses an unmanned vehicle distributed information safety protection device, and belongs to the field of vehicle information safety, and the device comprises a monitoring and early warning module which carries out the environment perception and state monitoring of a surrounding environment, and carries out the early warning of a generated abnormal operation behavior; the distributed protection module is used for realizing cooperative protection among the modules and dynamically adjusting a protection strategy; the self-destruction module is used for rating threats when uncontrollable threats are detected, automatically starting a self-destruction program and destroying sensitive data on the storage medium; a data module; the communication module is used for realizing information transmission between unmanned vehicle nodes; the comprehensive control module is used for organically combining the modules and carrying out centralized command and resource allocation; according to the unmanned vehicle distributed information safety protection device, the distributed protection module, the monitoring and early warning module, the self-destruction module and the communication module are combined, all-around information safety protection is provided for an unmanned vehicle system, data leakage is prevented, and the system is prevented from being maliciously controlled by others.

Description

Technical Field

[0001] The present invention relates to the field of vehicle information security technology, and in particular to a distributed information security protection device for unmanned vehicles. Background Art

[0002] With the advancement of technology, the application of unmanned vehicles (AVs) is becoming increasingly widespread across various fields. They can perform a variety of tasks, including field exploration, inspection, monitoring, and transportation. This not only improves work efficiency but also reduces labor costs and the risks of working in harsh conditions. However, in complex environments, AVs face even greater risks, particularly information security threats such as electromagnetic interference and hacker attacks. If the control and information systems of an AV are hacked or damaged, not only can the ongoing operation be halted or rendered ineffective, but it can also allow malicious actors to take control and carry out destructive activities. Therefore, ensuring the information security of AVs during their missions is crucial.

[0003] Traditional information security protection devices for unmanned vehicles currently focus on single-point protection, making them vulnerable to breaches. They also fail to consider the harmless processing of sensitive information in emergency situations, posing significant security risks and failing to fully meet safety requirements. To address these issues, a distributed information security protection device for unmanned vehicles was designed. Summary of the Invention

[0004] The purpose of the present invention is to provide a distributed information security protection device for unmanned vehicles to solve the problem that the existing technologies in the above-mentioned background technologies mostly focus on data encryption communication, firewalls and intrusion detection, but cannot implement effective protection measures when facing complex and high-intensity security threats, especially when facing high-intensity intrusions.

[0005] To achieve the above objectives, the present invention provides a distributed information security protection device for unmanned vehicles, comprising:

[0006] The monitoring and early warning module integrates multiple sensors to sense the surrounding environment and monitor its status. It also uses machine learning algorithms to identify and monitor system status in real time and issue early warnings for abnormal operating behaviors.

[0007] Distributed protection modules enable collaborative protection among modules, share security information, and dynamically adjust protection strategies;

[0008] The self-destruct module, when detecting an uncontrollable threat, rates the threat and automatically triggers a self-destruct program to destroy sensitive data on the storage medium;

[0009] Data module, including data storage unit, data erasure unit, data encryption and decryption unit and data redundancy unit;

[0010] Communication module, which realizes information transmission between unmanned vehicle nodes;

[0011] The integrated control module organically combines the above modules to carry out centralized command and resource allocation.

[0012] Preferably, the monitoring and warning module uploads the data collected by the sensors on the unmanned vehicle to the shared network in real time to form global situation awareness. The machine learning algorithm adopts an adaptive hybrid model machine learning algorithm, which is composed of a neural network for feature learning, Bayesian inference for capturing the uncertainty of input data, and an adaptive mechanism for dynamically adjusting the neural network parameters. The expression of the adaptive hybrid model is as follows:

[0013] P(y|x,0)=∫P(y|z,0)P(z|x,0)P(θ|D)dzdθ;

[0014] Where x is the input data; y is the output prediction; z is the latent variable; θ is the model parameter; D is the current training data set;

[0015] First, the features of the input data x are extracted through a deep neural network, and the latent variable distribution P(z|x,θ) is generated. The feature extraction formula is:

[0016] z=f NN (x;θ NN );

[0017] Among them, f NN is the forward propagation function of the neural network; θ NN are the parameters of the neural network, and then the uncertainty of the model parameters θ is captured by Bayesian inference, whose posterior distribution is:

[0018]

[0019] Where P(D|θ) is the likelihood function of the data; P(θ) is the prior distribution of the model parameters;

[0020] The final output distribution prediction model has the following output distribution:

[0021]

[0022] Through streaming big data analysis, the real-time uploaded data is processed and analyzed, and abnormal conditions are monitored and identified.

[0023] Preferably, streaming big data analysis and status monitoring includes three parts: data stream processing framework, status monitoring model, and adaptive threshold adjustment mechanism, specifically:

[0024] First, the data stream is segmented and processed. Assuming that the data stream arrives in the form of time series, it is recorded as Each x tis the data collected at time t; the data stream is sliced ​​by sliding window, as shown below:

[0025] W t ={x t-k+1 ,...,x t};

[0026] Among them, W t is the data window at time t; k is the window size;

[0027] In each window W t The behavior pattern of the data in the window is analyzed through the state monitoring model. If the data deviates from the normal behavior, it is identified as an abnormal state;

[0028] The condition monitoring model specifically includes:

[0029] average value:

[0030] Standard Deviation:

[0031] Abnormal judgment:

[0032] When S t >δ, where δ is the set threshold, it is considered that abnormal behavior exists at the current time point t;

[0033] The threshold δ is adaptively adjusted to adapt to the dynamic environment of the unmanned vehicle and the following formula is set:

[0034] δ t =α·σ t +β;

[0035] Among them, α and β are dynamic adjustment parameters; δ t is the standard deviation of the current window, which is used to dynamically reflect data changes;

[0036] Combined with the status monitoring results within the window, identify abnormal behavior and generate responses. t >δ t When t+1 is entered, the window is updated to W. t +1, repeat the above steps to continue early warning monitoring.

[0037] Preferably, in the distributed protection module, the TLS two-way authentication protocol is used between each node to protect the transmission of security information, risk identification data and protection strategies shared between nodes. When a node detects a threat, the system adopts a dynamic protection strategy through a collaborative mechanism, specifically:

[0038] The system monitors environmental threats in real time through the sensor data of the unmanned vehicle and evaluates the threat score S threat , the calculation formula is as follows:

[0039] S threat =α1·T impact +α2·T proximity +α3·T confidence ;

[0040] Among them, T impact is the potential severity of the threat; T proximity is the spatial proximity of the threat; T confidence is the confidence level of threat identification; α1α2α3 are weight factors;

[0041] Based on the threat score, the priority of each protection strategy is dynamically assigned to ensure that key strategies are executed first, which is expressed as:

[0042]

[0043] Among them, w i For strategy P i The weight of w j For strategy P i priority;

[0044] At the same time, according to the current resource status R avail And policy priority, resources are allocated to each protection strategy, expressed as:

[0045] R i =P i ·R avail ;

[0046] Among them, R i Assigned to strategy P i resources.

[0047] Preferably, in the distributed protection module, the sensor parameter T is adjusted according to the threat score. k , specifically:

[0048] LiDAR scanning frequency adjustment:

[0049] f lidar =min(f max ,β1·S threat );

[0050] Among them, f lidar is the laser radar scanning frequency; f max is the maximum scanning frequency of the laser radar; β1 is the adjustment coefficient;

[0051] Camera frequency adjustment:

[0052] f camera =min(f max ,β2·S threat );

[0053] Among them, f camera is the camera frame rate; β2 is the adjustment coefficient;

[0054] At each time step t real Update the threat score and protection policy. To update the threat score, follow the following formula:

[0055]

[0056] Among them, γ is the smoothing coefficient; Score the measurement threat for the current time step;

[0057] Reallocate priorities and resources based on the new scores, and activate the self-destruct module to self-destruct when the threat is high.

[0058] Preferably, the activation of the self-destruct module is based on threat detection, including external threat detection and internal threat detection. External threat detection detects abnormal behavior by monitoring the external laser radar and camera of the early warning module. Internal threat detection targets malicious tampering, violent disassembly or signal interference attacks, and uses the built-in acceleration sensor to monitor the internal status of the vehicle in real time. Specifically:

[0059] Accelerometers are used to detect abnormal environments and monitor for severe vibration or disassembly.

[0060]

[0061] When A(t)>A threshold When there is a threat of dismantling;

[0062] in, is the triaxial acceleration component; A threshold is the shock threshold;

[0063] By electronic shielding the signal strength P jam Detect whether the communication frequency band is interfered with. If P jam >P threshold is threatened by signal interference, where P threshold is the signal interference threshold;

[0064] Calculate the internal threat score S based on the above factors internal :

[0065] S internal =β1·A(t)+β2·P jam ;

[0066] Score of external threat threat and internal rating S internal Comprehensively evaluate whether to initiate the self-destruction program; the trigger mechanism uses logical judgment and preset thresholds. When the comprehensive threat score exceeds the threshold, the self-destruct module is triggered:

[0067] S total =γ1·S threat +γ2·S internal ;

[0068] Among them, S total is the comprehensive threat score; γ1 and γ2 are weight factors used to balance the impact of internal and external threats;

[0069] When S total >S threshold When the self-destruct module is activated, S threshold The threat score threshold that triggers self-destruction; data destruction is achieved by activating the self-destruct module, which is carried out by overwriting data, destroying storage chips, and encrypting and destroying key data;

[0070] Overwrite the storage medium multiple times to clear data traces, and use a pseudo-random number generator to generate overwrite data:

[0071] D fianl = Random(k,n);

[0072] Where k is the number of overwriting times and n is the size of the storage medium;

[0073] Send high-voltage electric pulses to the memory chip to physically destroy the chip. The pulse formula is as follows:

[0074] Vpulse=Vmax·sin(2πft);

[0075] Among them, V max is the maximum pulse voltage, and f is the pulse frequency.

[0076] Preferably, each unit of the data module is specifically:

[0077] The data storage unit uses a distributed database for data storage, including a hierarchical storage strategy and a role-based access control mechanism;

[0078] The data erasure unit provides logical erasure and physical erasure methods, overwriting random data multiple times or using disk fragmentation physical means, and configures an automated erasure process; it automatically executes data erasure operations when it detects that the storage medium is no longer secure or the data is no longer needed;

[0079] The data encryption and decryption unit uses the AES symmetric encryption algorithm and the RSA asymmetric encryption algorithm to encrypt and decrypt data, and integrates a TPM or HSM hardware encryption module, while configuring a secure key management system;

[0080] The data redundancy unit adopts a multi-copy strategy and an off-site backup strategy, ensures data consistency and high availability through a distributed consistency algorithm, and sets up an automatic recovery mechanism to automatically restore data from backup or redundant nodes when a data node failure is detected, ensuring the continuous availability of the system.

[0081] Preferably, the communication module adopts multiple methods to achieve communication security, including:

[0082] First, the communication content is encrypted using symmetric encryption and asymmetric encryption algorithms. The symmetric encryption process will use the Advanced Encryption Standard AES high-strength symmetric encryption algorithm to divide the input plaintext into blocks of fixed size and generate ciphertext through multiple rounds of encryption operations. Specifically: First, the plaintext is XORed with the initial key to obtain the initial state of encryption; then, each byte is nonlinearly replaced by S-Box through byte substitution; then row shift is performed to cyclically shift each row of the state matrix; then in the column obfuscation step, each column of data is obfuscated through matrix multiplication operations on a finite field; finally, the key plus round is used again to perform XOR operation with the subkey. The whole process is repeated 10 times, and the encrypted ciphertext is finally output. The AES decryption process is the reverse operation of encryption, and the plaintext is gradually restored through reverse processing. Specifically, first, the initial key is added to the ciphertext block, and the ciphertext is XORed with the subkey of the last round; then, inverse column confusion is performed, and the reverse finite field matrix multiplication operation is performed on each column of data; then, the reverse row shift shifts the rows of the state matrix back to their original positions according to the shift rule during encryption; then, in the inverse byte substitution, the inverse S-Box is used to perform nonlinear substitution on each byte to restore the original state. By repeating the above operations, the reverse processing is carried out round by round, and the plaintext data is finally restored.

[0083] The communication module uses RSA as an asymmetric encryption algorithm, encrypting data with a public key and decrypting data with a private key. Specifically, the key pair is first generated. Two large prime numbers p and q are randomly generated. n = p·q and the Euler function φ(n) = (p-1)(q-1) are calculated. The encryption exponent e is selected so that 1 < e < φ(n) and gcd(e, φ(n)) = 1. The decryption exponent d is calculated to satisfy the following formula:

[0084] e·d≡1(modφ(n));

[0085] Among them, the public key is (e,n) and the private key is (d,n);

[0086] Use the public key to the symmetric key Kδ Encryption is as follows:

[0087]

[0088] Among them, C k is the encrypted ciphertext;

[0089] Finally, the private key is used to restore the symmetric key as shown below:

[0090]

[0091] The communication module uses a multi-factor authentication mechanism to verify the identities of both communicating parties. At the same time, hash functions and digital signature technologies are used to ensure the integrity and authenticity of the communication content. SHA-256 is a one-way cryptographic hash function that maps input data of any length to a 256-bit hash value. Specifically, the input data is padded to make its length an integer multiple of 512 bits. Then, a 256-bit hash buffer is initialized, containing the initial values ​​of eight 32-bit words. Then, for each 512-bit block, the value of the hash buffer is updated through a series of bit operations, logical operations, and circular shift operations. Finally, the eight 32-bit words in the hash buffer are concatenated to form a 256-bit hash value as output. ECDSA is a digital signature algorithm based on elliptic curve cryptography. It generates a hash value H(M) by using the hash algorithm SHA-256 on the message M, and generates a signature pair (r, s) using the private key d and the random number k, as shown in the following formula:

[0092] r=(kG) x mod n;

[0093] s=k -1 (H(M)+dr)mod n;

[0094] Where G is the base point of the elliptic curve;

[0095] Verify the signature pair (r,s) using the public key Q = dG:

[0096] w=s -1 mod n;

[0097] μ1=H(M)w mod n,μ2=rwmod n;

[0098] (x,y)=μ1G+μ2Q;

[0099] If r≡x mod n, the signature is valid, otherwise the signature is invalid;

[0100] Finally, an encrypted communication tunnel is established using a secure network protocol.

[0101] Therefore, the present invention adopts the above-mentioned distributed information security protection device for unmanned vehicles, aiming to improve the information security and reliability of unmanned vehicles during the execution of tasks, so that they can effectively deal with complex and changeable security threats. An encrypted communication protocol is used between each node to share security information in real time, and the system has the ability to dynamically adjust the protection strategy. The various subsystems of the unmanned vehicle are modularly designed so that each protection module can work independently or be flexibly combined to meet diverse protection needs and have good scalability. By integrating self-destruct devices in modules such as lidar and cameras, when uncontrollable high-intensity threats are detected, the self-destruct program can be automatically triggered to completely destroy sensitive data. Modules such as cameras and lidars also have built-in multiple sensors and monitoring software, which use machine learning algorithms to monitor the system status in real time and quickly respond to abnormal behavior. Finally, encryption algorithms and authentication mechanisms are used to ensure that information transmission between the submodules of the unmanned vehicle and the command center nodes is safe and reliable.

[0102] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 This is a schematic structural diagram of a distributed information security protection device for an unmanned vehicle according to the present invention;

[0104] Figure 2 This is a flowchart of the workflow of a distributed information security protection device for unmanned vehicles of the present invention. DETAILED DESCRIPTION

[0105] The following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort shall fall within the scope of protection of the present invention.

[0106] See also Figure 1 , a distributed information security protection device for an unmanned vehicle, comprising:

[0107] The monitoring and early warning module integrates multiple sensors such as lidar, cameras, infrared sensors, and ultrasonic sensors to conduct all-round environmental perception and status monitoring of the surrounding environment. It uses an adaptive hybrid model machine learning algorithm that combines deep learning and Bayesian inference to dynamically adjust threat detection capabilities in complex systems. Nonlinear features are modeled through neural networks, and Bayesian inference is used to capture the uncertainty of the model to dynamically adjust the system's detection capabilities. Specifically, this model can adaptively adjust its parameters so that the model can remain robust when faced with changing input distributions. The core of the adaptive hybrid model can be divided into several parts: neural networks are used for feature learning, Bayesian inference is used to capture the uncertainty of input data, and adaptive mechanisms dynamically adjust the parameters of the neural network. The expression of the adaptive hybrid model is as follows:

[0108] P(y|x,0)=∫P(y|z,0)P(z|x,0)P(θ|D)dzdθ;

[0109] Where x is the input data (such as radar and camera data from an unmanned vehicle); y is the output prediction (such as whether there is a security threat); z is the latent variable (the hidden variable generated by the neural network); θ is the model parameter; and D is the current training dataset.

[0110] First, the features of the input data x are extracted through a deep neural network (DNN) and the latent variable distribution P(z|x,θ) is generated. The feature extraction formula is:

[0111] z=f NN (x;θ NN );

[0112] Among them, f NN is the forward propagation function of the neural network; θ NN are the parameters of the neural network, and then the uncertainty of the model parameters θ is captured by Bayesian inference, whose posterior distribution is:

[0113]

[0114] Where P(D|θ) is the likelihood function of the data; P(θ) is the prior distribution of the model parameters;

[0115] The final output distribution prediction model has the following output distribution:

[0116]

[0117] Streaming big data analysis processes and analyzes real-time uploaded data, and monitors and identifies abnormal conditions. Streaming big data analysis and status monitoring include three parts: data stream processing framework, status monitoring model, and adaptive threshold adjustment mechanism. Specifically:

[0118] First, the data stream is segmented and processed. Assuming that the data stream arrives in the form of time series, it is recorded as Each x t is the data collected at time t; the data stream is sliced ​​by sliding window, as shown below:

[0119] W t ={x t-k+1 ,...,x t};

[0120] Among them, W t is the data window at time t; k is the window size;

[0121] In each window W t The behavior pattern of the data in the window is analyzed through the state monitoring model. If the data deviates from the normal behavior, it is identified as an abnormal state;

[0122] The condition monitoring model specifically includes:

[0123] average value:

[0124] Standard Deviation:

[0125] Abnormal judgment:

[0126] When S t >δ, where δ is the set threshold, it is considered that abnormal behavior exists at the current time point t;

[0127] The threshold δ is adaptively adjusted to adapt to the dynamic environment of the unmanned vehicle and the following formula is set:

[0128] δ t =α·σ t +β;

[0129] Among them, α and β are dynamic adjustment parameters; δ t is the standard deviation of the current window, which is used to dynamically reflect data changes;

[0130] Combined with the status monitoring results within the window, identify abnormal behavior and generate responses. t >δ t When t+1 is entered, the window is updated to W. t +1, repeat the above steps to continue early warning monitoring.

[0131] The distributed protection module coordinates protection between the various subsystems within the unmanned vehicle, sharing security information and dynamically adjusting protection strategies to enhance the system's overall protection capabilities. By equipping the unmanned vehicle with independent protection units, including firewalls, intrusion detection systems (IDS), and antivirus software, it protects against network attacks and electronic interference. The vehicle's navigation radar and camera are reused as primary sensing devices to establish a security protection system. The TLS mutual authentication protocol is used between nodes, providing dual identity verification. Session keys are securely exchanged using asymmetric encryption and generated independently for each session, reducing the risk of key leakage. Within the distributed protection module, TLS is primarily used to protect the transmission of shared security information, threat identification data, and protection strategies between nodes, preventing eavesdropping, tampering, and spoofing attacks. To meet the low-latency, high-security, and resource-constrained requirements of the unmanned vehicle environment, the TLS 1.3 protocol and its 0-RTT handshake feature are used to reduce latency. Lightweight encryption algorithms (such as ChaCha20-Poly1305) and hardware acceleration technologies are used to improve encryption performance. Key management security is enhanced through session resumption and dynamic key update mechanisms. In distributed scenarios, multicast TLS and multipath transmission are used to optimize network adaptation, and dynamic certificate distribution and decentralized key management are combined to improve the efficiency of trust management of communication nodes. Through these optimizations, TLS can better meet the requirements of the distributed protection module of unmanned vehicles for efficient, secure, and low-latency communication.

[0132] In the distributed protection module, when a node detects a threat, the system implements dynamic protection strategies through a collaborative mechanism to enhance overall protection capabilities. First, the node immediately sends an alert to surrounding nodes and the cloud control system. By sharing the threat source's location, attributes, and severity, collaborative protection is triggered. Simultaneously, the vehicle's route is adjusted based on the threat situation, replanning to avoid threat areas and incorporating safety constraints to optimize the route. Regarding resource allocation, the system dynamically allocates computing power and communication bandwidth to prioritize threat-related tasks. Furthermore, sensor parameters are dynamically adjusted, such as increasing the scanning frequency of the lidar and the frame rate of the camera, or activating backup sensors (such as infrared sensors) to enhance perception capabilities, ensuring high-precision threat detection in complex environments. For higher threat levels, the system can activate electronic shielding devices for active jamming, forge GPS signals to confuse attackers, and notify surrounding command vehicles and the command center. In the face of severe threats, the emergency response mechanism is activated, activating the data module for data backup and the self-destruct module for self-destruction. All threat information is uploaded to the cloud control system in real time, which analyzes and optimizes the global protection strategy and allocates support nodes. These protection strategies are driven by threat scores and dynamically adjust priorities, resource allocation, and parameter settings to achieve flexible and adaptive protection capabilities and ensure the real-time responsiveness and safety of the unmanned vehicle system.

[0133] The core of dynamic adjustment of protection strategy lies in real-time perception of threat situation and flexible response. The system uses the sensor data (lidar and camera) of the unmanned vehicle to monitor environmental threats in real time and evaluate the threat score S threat , the calculation formula is as follows:

[0134] S threat =α1·T impact +α2·T proximity +α3·T confidence ;

[0135] Among them, T impact is the potential severity of the threat (range: [0, 1]); T proximity is the spatial proximity of the threat (range: [0, 1]); T confidence is the confidence level of threat identification (range: [0, 1]); α1, α2, and α3 are weight factors;

[0136] Based on the threat score, the priority of each protection strategy is dynamically assigned to ensure that key strategies are executed first, such as path adjustment and sensor optimization. This is expressed as:

[0137]

[0138] Among them, w i For strategy P i The weight of (preset according to the importance of the strategy); w j For strategy P i priority (range: [0,1]);

[0139] At the same time, according to the current resource status R avail And policy priority, resources are allocated to each protection strategy, expressed as:

[0140] R i =P i ·R avail ;

[0141] Among them, R i Assigned to strategy P i resources.

[0142] Adjust the sensor parameter T according to the threat score k , specifically:

[0143] LiDAR scanning frequency adjustment:

[0144] f lidar =min(f max ,β1·S threat );

[0145] Among them, f lidaris the laser radar scanning frequency; f max is the maximum scanning frequency of the laser radar; β1 is the adjustment coefficient;

[0146] Camera frequency adjustment:

[0147] f camera =min(f max ,β2·S threat );

[0148] Among them, f camera is the camera frame rate; β2 is the adjustment coefficient;

[0149] At each time step t real Update the threat score and protection policy. To update the threat score, follow the following formula:

[0150]

[0151] Among them, γ is the smoothing coefficient; Score the measurement threat for the current time step;

[0152] Reallocate priorities and resources based on the new scores, and activate the self-destruct module to self-destruct when the threat is high.

[0153] The self-destruct module is the last line of defense for ensuring data security and hardware protection in the unmanned vehicle system. When the protection system detects an uncontrollable threat, it rates the threat and automatically triggers the self-destruct program to completely destroy sensitive data on the storage medium, thereby preventing information leakage. The self-destruct module is activated based on threat detection. The monitoring and early warning module detects abnormal behavior. The distributed protection module classifies external threats and dynamically adjusts the protection strategy. When a high-intensity threat or uncontrollable security incident is detected, the self-destruct module activates the data destruction function through a trigger mechanism and responds quickly to prevent critical data leakage or malicious use of the device. This includes external threat detection and internal threat detection. External threat detection detects abnormal behavior through the monitoring and early warning module's external laser radar and camera. Internal threat detection targets malicious tampering, violent disassembly, or signal interference, and uses the built-in acceleration sensor to monitor the vehicle's internal status in real time. Specifically:

[0154] Accelerometers are used to detect abnormal environments and monitor for severe vibration or disassembly.

[0155]

[0156] When A(t)>A threshold When there is a threat of dismantling;

[0157] in, is the triaxial acceleration component; A threshold is the shock threshold;

[0158] By electronic shielding the signal strength P jam Detect whether the communication frequency band is interfered with. If P jam >P threshold is threatened by signal interference, where P threshold is the signal interference threshold;

[0159] Calculate the internal threat score S based on the above factors internal :

[0160] S internal =β1·A(t)+β2·P jam ;

[0161] Score of external threat threat and internal rating S internal Comprehensively evaluate whether to initiate the self-destruction program; the trigger mechanism uses logical judgment and preset thresholds. When the comprehensive threat score exceeds the threshold, the self-destruct module is triggered:

[0162] S total =γ1·S threat +γ2·S internal ;

[0163] Among them, S total is the comprehensive threat score; γ1 and γ2 are weight factors used to balance the impact of internal and external threats;

[0164] When S total >S threshold When the self-destruct module is activated, S threshold The threat score threshold that triggers self-destruction. By activating the self-destruct module, data destruction is achieved to prevent the unauthorized access of sensitive data stored in the autonomous vehicle. Data destruction is primarily carried out through data overwriting, storage chip destruction, and encryption and destruction of critical data.

[0165] Overwrite the storage medium multiple times to clear data traces, and use a pseudo-random number generator to generate overwrite data:

[0166] D fianl = Random(k,n);

[0167] Where k is the number of overwriting times and n is the size of the storage medium;

[0168] Send high-voltage electric pulses to the memory chip to physically destroy the chip. The pulse formula is as follows:

[0169] Vpulse=Vmax·sin(2πft);

[0170] Among them, V max is the maximum pulse voltage, and f is the pulse frequency.

[0171] The data module mainly includes a data storage unit, a data erasure unit, a data encryption and decryption unit, and a data redundancy unit;

[0172] The data storage unit uses a distributed database for data storage, and ensures high availability, integrity and legal access to data through hierarchical storage strategies and role-based access control (RBAC) mechanisms.

[0173] The data erasure unit provides logical erasure and physical erasure methods. It ensures that data cannot be recovered by repeatedly writing random data over or using physical means such as disk fragmentation. It also configures an automated erasure process to automatically perform data erasure operations when it detects that the storage medium is no longer safe or the data is no longer needed.

[0174] The data encryption and decryption unit uses symmetric encryption algorithms such as AES and asymmetric encryption algorithms such as RSA to encrypt and decrypt data to ensure the confidentiality and privacy of the data. It improves efficiency and security by integrating hardware encryption modules such as TPM or HSM, and configures a secure key management system (such as KMS) to ensure the safe storage and use of keys.

[0175] The data redundancy unit employs multiple replica strategies (such as three replicas) and off-site backup strategies, ensuring data consistency and high availability through a distributed consistency algorithm. It also incorporates an automatic recovery mechanism that automatically restores data from backup or redundant nodes when a data node failure is detected, ensuring continuous system availability. Through these functions and specific implementations, the data module provides comprehensive security within the distributed protection system for unmanned vehicle information security, effectively ensuring data confidentiality, integrity, and high availability.

[0176] Data storage and backup, data encryption and decryption, permissions management, data integrity checking, redundancy and fault tolerance, and data destruction are crucial for ensuring data security and integrity. Data storage and backup utilize a distributed file system to store data across multiple nodes and configure regular or real-time backup strategies to back up critical data to remote or cloud storage.

[0177] In the data module, data encryption and decryption are core components to ensure data security and integrity. High-strength encryption algorithms protect data, preventing theft or tampering of sensitive information during storage or transmission. Permission management, through role-based access control and multi-factor authentication, ensures that only authorized entities can access data. Data integrity checks combine secure hashing algorithms with blockchain technology to provide comprehensive data protection. The system uses Hyperledger Fabric as the blockchain framework, leveraging its high performance, access control, and privacy protection to log data operations. Each data operation (such as addition, modification, or deletion) generates a unique checksum using the SHA-256 hash algorithm. This checksum, along with information such as the operation type and timestamp, is packaged into a block, verified using the PBFT consensus algorithm, and then added to the blockchain. The blockchain's chain-like structure and distributed storage ensure the immutability of the data log; tampering with any log entry invalidates the entire chain. Furthermore, the system supports tracing back historical data operations against the blockchain ledger, verifying that the current data checksum matches the blockchain record, thereby detecting data tampering. This mechanism, combined with data operation rules defined by smart contracts, not only ensures data integrity and immutability, but also enhances transparency and traceability of data operations. Redundancy and fault tolerance ensure continuous data availability in the event of storage node failures by configuring data replication strategies and automatic fault recovery mechanisms. Data destruction triggers a self-destruct mechanism upon detecting a serious security threat, quickly and securely destroying sensitive data on the storage media through hardware meltdown or disk erasure. Through these methods, the data storage module effectively safeguards the security, integrity, and availability of autonomous vehicle data, ensuring that data can be securely stored and managed in a variety of complex environments and potential threats. Furthermore, critical data encryption keys can be directly cleared, rendering the data inaccessible. Data destruction effectively protects the autonomous vehicle's sensitive data and system security, preventing malicious exploitation.

[0178] The communication module employs multiple methods to ensure communication security and prevent interception. First, communication content is encrypted using symmetric encryption algorithms (such as AES) and asymmetric encryption algorithms (such as RSA) to ensure data confidentiality. The AES encryption process utilizes the high-strength symmetric encryption algorithm (AES (Advanced Encryption Standard)) to divide the input plaintext into fixed-size blocks (128 bits) and generate ciphertext through multiple rounds of encryption. First, the plaintext is XORed with the initial key to generate the initial encrypted state. Next, each byte is nonlinearly replaced using an S-Box through byte substitution. Row shifting is then performed to cyclically shift the rows of the state matrix. In the column obfuscation step, each column is obfuscated using matrix multiplication over a finite field. Finally, the key is XORed with the subkey again. This entire process is repeated 10 times, ultimately outputting the encrypted ciphertext. Each round leverages both linear and nonlinear transformations, making AES encryption highly secure. The AES decryption process is the reverse of encryption, gradually recovering the plaintext. First, the ciphertext block is rounded with the initial key, and the ciphertext is XORed with the subkey from the last round. Next, inverse column obfuscation is performed, inverting the finite field matrix multiplication operation on each column. Next, inverse row shifting shifts the rows of the state matrix back to their original positions according to the shifting rules used during encryption. Finally, in inverse byte substitution, each byte is nonlinearly replaced using the inverse S-Box to restore the original state. By repeating these steps, reversing the process round by round, the plaintext data is ultimately recovered. During decryption, the encryption steps are strictly followed in reverse order to ensure data consistency and integrity.

[0179] The communication module uses RSA as an asymmetric encryption algorithm, encrypting data with a public key and decrypting data with a private key. Specifically, the key pair is first generated. Two large prime numbers p and q are randomly generated. n = p·q and the Euler function φ(n) = (p-1)(q-1) are calculated. The encryption exponent e is selected so that 1 < e < φ(n) and gcd(e, φ(n)) = 1. The decryption exponent d is calculated to satisfy the following formula:

[0180] e·d≡1(modφ(n));

[0181] Among them, the public key is (e,n) and the private key is (d,n);

[0182] Use the public key to the symmetric key K δ Encryption is as follows:

[0183]

[0184] Among them, C k is the encrypted ciphertext;

[0185] Finally, the private key is used to restore the symmetric key as shown below:

[0186]

[0187] The communication module utilizes multi-factor authentication mechanisms (such as digital certificates and biometrics) to verify the identities of both communicating parties and prevent unauthorized access. Furthermore, hash functions and digital signature technologies ensure the integrity and authenticity of the communication content, preventing data tampering during transmission. SHA-256 is a one-way cryptographic hash function that maps input data of arbitrary length to a 256-bit hash value. First, the input data is padded to an integer multiple of 512 bits. Next, a 256-bit hash buffer is initialized with eight 32-bit words as initial values. For each 512-bit block, a series of bitwise operations, logical operations, and circular shifts are performed to update the hash buffer. Finally, the eight 32-bit words in the hash buffer are concatenated to form a 256-bit hash value as output. SHA-256's one-way nature, collision resistance, and sensitivity make it widely used in security applications such as data integrity verification and digital signatures. Its one-way nature makes it difficult to deducing the original data from the hash value. Its collision resistance ensures that the probability of finding two different inputs producing the same hash value is extremely low, and its sensitivity ensures that any slight change in the data will cause the hash value to change significantly. ECDSA (Elliptic Curve Digital Signature Algorithm) is a digital signature algorithm based on elliptic curve cryptography, which uses the computational complexity of the elliptic curve discrete logarithm problem to ensure security. By using the hash algorithm SHA-256 on the message M to generate the hash value H(M), the private key d and the random number k are used to generate the signature pair (r, s), as shown in the following formula:

[0188] r=(kG) x mod n;

[0189] s=k -1 (H(M)+dr)mod n;

[0190] Where G is the base point of the elliptic curve;

[0191] Verify the signature pair (r,s) using the public key Q = dG:

[0192] w=s -1 mod n;

[0193] μ1=H(M)w mod n,μ2=rw mod n;

[0194] (x,y)=μ1G+μ2Q;

[0195] If r≡x mod n, the signature is valid, otherwise the signature is invalid;

[0196] Finally, secure network protocols (such as TLS / SSL and VPN) are used to establish encrypted communication tunnels to further prevent data from being intercepted and tampered with. Through these measures, the communication module effectively protects the communication security of the unmanned vehicle system and ensures the confidentiality, integrity, and authenticity of data during transmission.

[0197] The integrated control module plays a core coordination and management role in the information security protection of unmanned vehicles, responsible for unified scheduling, policy management, and real-time response for the overall security architecture. It organically integrates the various security function modules of the unmanned vehicle (including the monitoring and warning module, distributed protection module, data module, communication module, and self-destruction module) to form an efficient closed-loop security system. For the monitoring and warning module, the integrated control module is responsible for aggregating threat intelligence, dynamically adjusting monitoring scope and frequency, and triggering response strategies when warning signals appear. For the distributed protection module, it manages node status and policy allocation, dispatches defense resources in real time to intercept threats, and integrates protection logs to optimize defense strategies. For the data module, it is responsible for unified management of storage, encryption, and backup operations, ensuring data integrity and rapid recovery capabilities, while cooperating with the communication module to ensure data transmission security. For the communication module, it dynamically adjusts encryption strategies, authenticates the identities of both communicating parties, and instructs the communication module to terminate or switch communication channels in the event of an anomaly. For the self-destruction module, it intelligently determines trigger conditions based on the threat level and, when necessary, destroys data and hardware to prevent the leakage of core information. Through the organic integration and precise scheduling of various modules, the integrated control module ensures that the unmanned vehicle has the ability to respond quickly, protect safety, manage data, and protect itself, achieving efficient, stable, and comprehensive information security protection in a complex operating environment.

[0198] Figure 2This is a workflow diagram for a distributed information security protection device for unmanned vehicles according to this embodiment. Specifically, the device includes: a monitoring and early warning module that uploads data collected by the unmanned vehicle's lidar and camera to a shared network in real time for global situational awareness. Using an adaptive hybrid model machine learning algorithm combining deep learning and Bayesian inference, the device processes the data through streaming big data analysis to detect anomalies. Identified threats are assessed, and if they are not, data is stored and securely communicated. If they are considered dangerous, the device classifies them into different levels, dynamically adjusts security protection strategies for low-threat situations, logs them, and securely communicates. Based on these, a warning signal is issued to notify surrounding vehicles and the command center. The distributed protection module shares security information between nodes, is equipped with firewalls and other protection units, and uses the TLS mutual authentication protocol to ensure communication security. It dynamically adjusts protection strategies based on the threat situation. When faced with a high threat, the self-destruct module is activated. Building on the work of the monitoring and early warning and distributed protection modules, the self-destruct module performs internal and external threat detection. When the comprehensive threat score exceeds a threshold, it activates the emergency response mechanism. The data module backs up data and transmits it to the cloud. The data is then destroyed by sending high-voltage pulses to the storage chip, physically damaging the chip or overwriting the data. The data module is responsible for data storage and backup, encryption and decryption, permissions management, integrity checking, redundancy fault tolerance, and destruction, utilizing a variety of technologies to ensure data security. The communication module utilizes symmetric and asymmetric encryption algorithms, multi-factor authentication, hash functions, and digital signature technologies, combined with secure network protocols, to ensure the confidentiality, integrity, and authenticity of communications. The integrated control module coordinates all modules, performing real-time monitoring, security policy management, and resource scheduling, forming an efficient and secure closed-loop system to ensure efficient, stable, and comprehensive information security protection for unmanned vehicles in complex environments. Through this process, a comprehensive overview of the unmanned vehicle's information security protection has been developed, significantly improving the system's safety and reliability.

[0199] Therefore, the present invention adopts the above-mentioned distributed information security protection device for unmanned vehicles, and provides all-round information security protection for the unmanned vehicle system by combining the distributed protection module, monitoring and early warning module, self-destruction module and communication module to prevent data leakage and the system from being maliciously controlled by others.

[0200] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A distributed information security protection device for unmanned vehicles, characterized in that: include: The monitoring and early warning module integrates multiple sensors to sense the surrounding environment and monitor its status. It also uses machine learning algorithms to identify and monitor system status in real time and issue early warnings for abnormal operating behaviors. Distributed protection modules enable collaborative protection among modules, share security information, and dynamically adjust protection strategies; The self-destruct module, when detecting an uncontrollable threat, rates the threat and automatically triggers a self-destruct program to destroy sensitive data on the storage medium; Data module, including data storage unit, data erasure unit, data encryption and decryption unit and data redundancy unit; Communication module, which realizes information transmission between unmanned vehicle nodes; The integrated control module organically combines the above modules to carry out centralized command and resource allocation.

2. The distributed information security protection device for unmanned vehicles according to claim 1, characterized in that: The monitoring and warning module uploads data collected by sensors on the unmanned vehicle to a shared network in real time to form global situational awareness. The machine learning algorithm adopts an adaptive hybrid model machine learning algorithm, which consists of a neural network for feature learning, Bayesian inference for capturing the uncertainty of input data, and an adaptive mechanism for dynamically adjusting the neural network parameters. The expression of the adaptive hybrid model is as follows: P(y|x,0)=∫P(y|z,0)P(z|x,0)P(θ|D)dzdθ; Where x is the input data; y is the output prediction; z is the latent variable; θ is the model parameter; D is the current training data set; First, the features of the input data x are extracted through a deep neural network, and the latent variable distribution P(z|x,θ) is generated. The feature extraction formula is: z=f NN (x;θ NN ); Among them, f NN is the forward propagation function of the neural network; θ NN are the parameters of the neural network, and then the uncertainty of the model parameters θ is captured by Bayesian inference, whose posterior distribution is: P(θ|D)∝P(D|θ)P(θ); Where P(D|θ) is the likelihood function of the data; P(θ) is the prior distribution of the model parameters; The final output distribution prediction model has the following output distribution: Through streaming big data analysis, the real-time uploaded data is processed and analyzed, and abnormal conditions are monitored and identified.

3. The distributed information security protection device for unmanned vehicles according to claim 2, characterized in that: Streaming big data analysis and status monitoring includes three parts: data stream processing framework, status monitoring model, and adaptive threshold adjustment mechanism. Specifically: First, the data stream is segmented and processed. Assuming that the data stream arrives in the form of time series, it is recorded as Each x t is the data collected at time t; the data stream is sliced ​​by sliding window, as shown below: W t ={x t-k+1 ,...,x t }; Among them, W t is the data window at time t; k is the window size; In each window W t The behavior pattern of the data in the window is analyzed through the state monitoring model. If the data deviates from the normal behavior, it is identified as an abnormal state; The condition monitoring model specifically includes: average value: Standard Deviation: Abnormal judgment: When S t >δ, where δ is the set threshold, it is considered that abnormal behavior exists at the current time point t; The threshold δ is adaptively adjusted to adapt to the dynamic environment of the unmanned vehicle and the following formula is set: d t =a·s t +b; Among them, α and β are dynamic adjustment parameters; δ t is the standard deviation of the current window, which is used to dynamically reflect data changes; Combined with the status monitoring results within the window, identify abnormal behavior and generate responses. t >δ t When t+1 is entered, the window is updated to W. t +1, repeat the above steps to continue early warning monitoring.

4. The distributed information security protection device for unmanned vehicles according to claim 1, characterized in that: In the distributed protection module, the TLS two-way authentication protocol is used between nodes to protect the transmission of shared security information, threat identification data, and protection strategies. When a node detects a threat, the system adopts a dynamic protection strategy through a collaborative mechanism. Specifically: The system monitors environmental threats in real time through the sensor data of the unmanned vehicle and evaluates the threat score S threat , the calculation formula is as follows: S threat =α1·T impact +α2·T proximity +α3·T confidence ; Among them, T impact is the potential severity of the threat; T proximity is the spatial proximity of the threat; T confidence is the confidence level of threat identification; α1, α2, and α3 are all weight factors; Based on the threat score, the priority of each protection strategy is dynamically assigned to ensure that key strategies are executed first, which is expressed as: Among them, w i For strategy P i The weight of w j For strategy P i priority; At the same time, according to the current resource status R avail And policy priority, resources are allocated to each protection strategy, expressed as: R i =P i ·R avail ; Among them, R i Assigned to strategy P i resources.

5. The distributed information security protection device for unmanned vehicles according to claim 4 is characterized in that: In the distributed protection module, it also includes adjusting the sensor parameter T according to the threat score k , specifically: LiDAR scanning frequency adjustment: f lidar =min(f max ,β1·S threat ); Among them, f lidar is the laser radar scanning frequency; f max is the maximum scanning frequency of the laser radar; β1 is the adjustment coefficient; Camera frequency adjustment: f camera =min(f max ,β2·S threat ); Among them, f camera is the camera frame rate; β2 is the adjustment coefficient; At each time step t real Update the threat score and protection policy. To update the threat score, follow the following formula: Among them, γ is the smoothing coefficient; Score the measurement threat for the current time step; Reallocate priorities and resources based on the new scores, and activate the self-destruct module to self-destruct when the threat is high.

6. The distributed information security protection device for unmanned vehicles according to claim 5, characterized in that: The self-destruct module is activated based on threat detection, including external threat detection and internal threat detection. External threat detection uses the monitoring early warning module's external lidar and camera to detect abnormal behavior. Internal threat detection targets malicious tampering, violent disassembly, or signal interference attacks, using the built-in acceleration sensor to monitor the vehicle's internal status in real time. Specifically: Accelerometers are used to detect abnormal environments and monitor for severe vibration or disassembly. When A(t)>A threshold When there is a threat of dismantling; in, is the triaxial acceleration component; A threshold is the shock threshold; By electronic shielding the signal strength P jam Detect whether the communication frequency band is interfered with. If P jam >P threshold is threatened by signal interference, where P threshold is the signal interference threshold; Calculate the internal threat score S based on the above factors internal : S internal =β1·A(t)+β2·P jam ; Score of external threat threat and internal rating S internal Comprehensively evaluate whether to initiate the self-destruction program; the trigger mechanism uses logical judgment and preset thresholds. When the comprehensive threat score exceeds the threshold, the self-destruct module is triggered: S total =γ1·S threat +γ2·S internal ; Among them, S total is the comprehensive threat score; γ1 and γ2 are weight factors used to balance the impact of internal and external threats; When S total >S threshold When the self-destruct module is activated, S threshold The threat score threshold that triggers self-destruction; data destruction is achieved by activating the self-destruct module, which is carried out by overwriting data, destroying storage chips, and encrypting and destroying key data; Overwrite the storage medium multiple times to clear data traces, and use a pseudo-random number generator to generate overwrite data: D fianl =Random(k,n); Where k is the number of overwriting times and n is the size of the storage medium; Send high-voltage electric pulses to the memory chip to physically destroy the chip. The pulse formula is as follows: Vpulse=Vmax·sin(2πft); Among them, V max is the maximum pulse voltage, and f is the pulse frequency.

7. The distributed information security protection device for unmanned vehicles according to claim 1, characterized in that: The specific units of the data module are: The data storage unit uses a distributed database for data storage, including a hierarchical storage strategy and a role-based access control mechanism; The data erasure unit provides logical erasure and physical erasure methods, overwriting random data multiple times or using physical means of disk fragmentation, and configures an automated erasure process; Automatically perform data erasure operations when it detects that the storage medium is no longer secure or the data is no longer needed; The data encryption and decryption unit uses the AES symmetric encryption algorithm and the RSA asymmetric encryption algorithm to encrypt and decrypt data, and integrates a TPM or HSM hardware encryption module, while configuring a secure key management system; The data redundancy unit adopts a multi-copy strategy and an off-site backup strategy, ensures data consistency and high availability through a distributed consistency algorithm, and sets up an automatic recovery mechanism to automatically restore data from backup or redundant nodes when a data node failure is detected, ensuring the continuous availability of the system.

8. The distributed information security protection device for unmanned vehicles according to claim 1, characterized in that: The communication module uses a variety of methods to achieve communication security, including: First, the communication content is encrypted using symmetric encryption and asymmetric encryption algorithms. The symmetric encryption process will use the Advanced Encryption Standard AES high-strength symmetric encryption algorithm to divide the input plaintext into blocks of fixed size and generate ciphertext through multiple rounds of encryption operations. Specifically: First, the plaintext is XORed with the initial key to obtain the initial state of encryption; then, each byte is nonlinearly replaced by S-Box through byte substitution; then row shift is performed to cyclically shift each row of the state matrix; then in the column obfuscation step, each column of data is obfuscated through matrix multiplication operations on a finite field; finally, the key plus round is used again to perform XOR operation with the subkey. The whole process is repeated 10 times, and the encrypted ciphertext is finally output. The AES decryption process is the reverse operation of encryption, and the plaintext is gradually restored through reverse processing. Specifically, first, the initial key is added to the ciphertext block, and the ciphertext is XORed with the subkey of the last round; then, inverse column confusion is performed, and the reverse finite field matrix multiplication operation is performed on each column of data; then, the reverse row shift shifts the rows of the state matrix back to their original positions according to the shift rule during encryption; then, in the inverse byte substitution, the inverse S-Box is used to perform nonlinear substitution on each byte to restore the original state. By repeating the above operations, the reverse processing is carried out round by round, and the plaintext data is finally restored. The communication module uses RSA as an asymmetric encryption algorithm, encrypting data with a public key and decrypting data with a private key. Specifically, the key pair is first generated. Two large prime numbers p and q are randomly generated. n = p·q and the Euler function φ(n) = (p-1)(q-1) are calculated. The encryption exponent e is selected so that 1 < e < φ(n) and gcd(e, φ(n)) = 1. The decryption exponent d is calculated to satisfy the following formula: e·d≡1(modφ(n)); Among them, the public key is (e,n) and the private key is (d,n); Use the public key to the symmetric key K δ Encryption is as follows: Among them, C k is the encrypted ciphertext; Finally, the private key is used to restore the symmetric key as shown below: The communication module uses a multi-factor authentication mechanism to verify the identities of both communicating parties. At the same time, hash functions and digital signature technologies are used to ensure the integrity and authenticity of the communication content. SHA-256 is a one-way cryptographic hash function that maps input data of any length to a 256-bit hash value. Specifically, the input data is padded to make its length an integer multiple of 512 bits. Then, a 256-bit hash buffer is initialized, containing the initial values ​​of eight 32-bit words. Then, for each 512-bit block, the value of the hash buffer is updated through a series of bit operations, logical operations, and circular shift operations. Finally, the eight 32-bit words in the hash buffer are concatenated to form a 256-bit hash value as output. ECDSA is a digital signature algorithm based on elliptic curve cryptography. It generates a hash value H(M) by using the hash algorithm SHA-256 on the message M, and generates a signature pair (r, s) using the private key d and the random number k, as shown in the following formula: r=(kG) x mod n; s=k -1 (H(M)+dr)mod n; Where G is the base point of the elliptic curve; Verify the signature pair (r,s) using the public key Q = dG: w=s -1 mod n; μ1=H(M)w mod n,μ2=rw mod n; (x,y)=μ1G+μ2Q; If r≡x mod n, the signature is valid, otherwise the signature is invalid; Finally, an encrypted communication tunnel is established using a secure network protocol.

Images