CN114726636B - Attack dynamic detection and identification method for heterogeneous cross-domain system - Google Patents
Attack dynamic detection and identification method for heterogeneous cross-domain system Download PDFInfo
- Publication number
- CN114726636B CN114726636B CN202210408479.7A CN202210408479A CN114726636B CN 114726636 B CN114726636 B CN 114726636B CN 202210408479 A CN202210408479 A CN 202210408479A CN 114726636 B CN114726636 B CN 114726636B
- Authority
- CN
- China
- Prior art keywords
- attack
- network
- domain
- heterogeneous cross
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
- H04L63/1416—Event detection, e.g. attack signature detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Hardware Design (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to an attack dynamic detection and identification method design containing a heterogeneous cross-domain system. The invention discloses an attack dynamic detection and identification method of a heterogeneous cross-domain system, which comprises the heterogeneous cross-domain system under network attack, a heterogeneous cross-domain system multi-source information fusion mechanism, a generation network G module, a discrimination network D module, a heterogeneous cross-domain system distributed fusion dynamic detection mechanism and an attack information on-line fusion identification mechanism. Aiming at the heterogeneous cross-domain system suffering from network attack, the invention designs a multi-source information fusion mechanism of the heterogeneous cross-domain system. Secondly, by combining the characteristics of methods for generating an anti-network iterative game and generating an anti, the G module of the generation network is used for carrying out dynamic fusion detection on the incompleteness, the asynchronism and the false performance of the measured information in the shortest time, and meanwhile, the D module of the discrimination network is used for training state data under network attack to carry out online fusion identification on attack information, so that distributed dynamic fusion detection of the state of the heterogeneous cross-domain system under attack and online fusion identification of the attack information are realized. The method has the advantages of solving the problem of measurement information failure of the heterogeneous cross-domain system caused by network attack and improving the safety of the heterogeneous cross-domain system.
Description
Technical Field
The invention relates to attack dynamic detection and identification, in particular to an attack dynamic detection and identification method of a heterogeneous cross-domain system.
Background
Under the promotion of the innovation of artificial intelligence technology, the heterogeneous cross-domain system relating to the heterogeneous outside the domain and the isomorphism inside the domain obtains a rapid development in the directions of intellectualization, autonomy and grouping, and is one of the future key research fields. According to the actual demand and the network information flow characteristics, the structure of the heterogeneous cross-domain system domain unit can be divided into an interactive control layer, an information fusion network layer and a multi-sensor cooperative sensing layer. In the heterogeneous cross-domain system, the network attack approach is more diversified, concealed and complicated. The network attack can not only occur in the network layer of information fusion, but also enter the interactive control layer of the cross-domain system to attack the controller and destroy the sensor in the multi-sensor cooperative sensing layer, so as to cause the problem of measurement information failure, and lead the unmanned system to generate behaviors such as situation misjudgment, decision error, behavior conflict and even system breakdown.
In order to deal with the problem of measurement information failure and improve the safety and reliability of the Unmanned Aerial Vehicle system, a plurality of signal Detection and Identification methods are proposed, and the problem of Unmanned Aerial Vehicle system state measurement information failure caused by network attack is analyzed in documents [ v.sadhu, s.zonouz and d.poipili, "On-board Deep-learning-based Unmanned Aerial Vehicle Fault Cause Detection and Identification,"2020IEEE International Conference On Robotics and Automation (ICRA), 2020, pp.5255-5261], and a new architecture based On a Deep convolution and long-term memory short-term neural network is proposed to realize the signal data Detection and Identification of the Unmanned Aerial Vehicle system so as to improve the safety of the Unmanned Aerial Vehicle system. A Subspace Tracking algorithm is researched in documents (Y.He, Y.Peng, S.Wang and D.Liu, 'ADMOST: UAV Flight Data analysis Detection and simulation via Online Subspace Tracking,' IEEE Transactions on Instrumentation and Measurement,2019,68 (4): 1035-1044) to realize the Detection and identification of the abnormal state of the unmanned aerial vehicle system and realize the stability and safety control of the unmanned aerial vehicle system.
The documents provide valuable research ideas for the research of the detection and identification of the abnormal state signals of the intelligent system. However, the research on detection and identification of attack signals of a heterogeneous cross-domain system is in a starting stage at home and abroad, and only a certain domain unit in the cross-domain system is researched in the existing documents and reports, such as an unmanned aerial vehicle system, so that the research on the attack dynamic detection and identification technology of the heterogeneous cross-domain system under network attack is very lacking. The heterogeneous cross-domain system comprising a plurality of domain units is an important research field of future intellectualization, grouping and autonomy and has strategic significance in the civil and military fields. Network attacks directly threaten the security and reliability of heterogeneous cross-domain systems. Therefore, the attack dynamic detection and identification method of the heterogeneous cross-domain system is a key technical problem which needs to be solved urgently in the development process of the heterogeneous cross-domain system.
Disclosure of Invention
The invention aims to solve the technical problem of state measurement information failure possibly occurring in a heterogeneous cross-domain system under network attack, and provides an attack dynamic detection and identification method for the heterogeneous cross-domain system.
In order to achieve the above object, according to an aspect of the embodiments of the present invention, a heterogeneous cross-domain system under a network attack including multiple domain units, a heterogeneous cross-domain system multi-source information fusion mechanism, a generation network G module, a discrimination network D module, a heterogeneous cross-domain system distributed fusion dynamic detection mechanism, and an attack information online fusion identification mechanism are provided.
Designing a heterogeneous cross-domain system multi-source information fusion mechanism by combining domain unit state characteristics and environment characteristics of the heterogeneous cross-domain system under network attack; by combining the characteristics of methods for generating a countermeasure network iterative game and generating countermeasures, the G module of the generation network is utilized to carry out dynamic fusion detection on the incompleteness, asynchronism and false performance of the measurement information in the shortest time; and performing online fusion identification on attack information by combining state data under the training attack of the discrimination network D module.
Specifically, the heterogeneous cross-domain system under the network attack can be divided into domain units U 1 Domain unit U 2 8230the domain unit U n . Wherein, the domain unit U 1 Including status signature data U 1s (x 1s1 ,x 1s2 8230quickly) and environmental characteristic data U 1e (x 1e1 ,x 1e2 8230in which x 1s1 ，x 1s2 Representation domain unit U 1 Characteristic of state of (1), x 1e1 ，x 1e2 Representation domain unit U 1 An environmental characteristic; domain unit U 2 Including status signature data U 2s (x 2s1 ,x 2s2 8230;) and environmental characteristic data U 2e (x 2e1 ,x 2e2 8230), where x 2s1 ，x 2s2 Presentation Domain Unit U 2 Status feature, x 2e1 ，x 2e2 Presentation Domain Unit U 2 An environmental characteristic; domain unit U n Including status signature data U ns (x ns1 ,x ns2 8230quickly) and environmental characteristic data U ne (x ne1 ,x ne2 ,…)，x ns1 ，x ns2 Representation domain unit U n Status feature, x ne1 ，x ne2 Presentation Domain Unit U n And (6) state characteristics.
Specifically, a heterogeneous cross-domain system domain unit U under network attack is analyzed 1 Domain unit U 2 8230the domain unit U n The state characteristic data and the environment characteristic data of the network are used for establishing a heterogeneous cross-domain system multi-source information fusion mechanism and generating real multi-source fusion state data M (U) under network attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne )；
Specifically, the generation network G module is based on a heterogeneous cross-domain system multi-source information fusion mechanism and is used for generating real multi-source fusion data M (U) under network attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) The method is obtained by randomly selecting variables as input and training. The discrimination network D module is real state data M (U) under attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And generating data for network output
As input, obtained by training. Wherein +>
Domain units U generated for training respectively 1 ，U 2 ，…,U n Status characteristic data of (a), based on the status of the device>
Domain units U generated for training respectively 1 ，U 2 ，…,U n The environmental characteristic data of (1).
In particular, based on raw materialsA network G and heterogeneous cross-domain system multi-source information fusion mechanism generates a group of data through training
Will generate data>
And true data M (U) under attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And repeatedly and iteratively comparing to realize dynamic fusion detection of imperfection, asynchronism and false of the measurement information in the shortest time.
Furthermore, the real state data M (U) under the attack is utilized 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And generating data
Training a discrimination network D, performing iterative game and mutual antagonistic training simultaneously through the generation network G and the discrimination network D, wherein the optimization target function is->
D and G respectively represent data sequences of the generation network and the judgment network, and the data sequences come from the actual sensing data distribution of the system. Finally reach nash equilibrium, thus realize the online fusion recognition of the attack information, receive the true compound attack information A (a) 1 ,a 2 8230; e.g. a) 1 Representing a double lasso attack, a 2 Indicating a spoofing attack.
The method has the advantages that the problem that the state measurement information of the heterogeneous cross-domain system fails under network attack can be effectively solved, so that the safety and reliability of the heterogeneous cross-domain system are improved, the attack resistance capability of the heterogeneous cross-domain system is enhanced, and the comprehensive safety guarantee capability of the system is improved.
The invention will be further explained with reference to the following description and embodiments in conjunction with the accompanying drawings.
Drawings
FIG. 1 is a schematic structural view of an embodiment of the present invention;
fig. 2 is a diagram of a generation countermeasure network architecture.
Detailed Description
In the following, a heterogeneous cross-domain system under network attack is taken as an example, and the technical solution of the present invention is described in detail, clearly and completely with reference to the accompanying drawings, so as to facilitate those skilled in the art to better understand the present invention.
Examples
Fig. 1 is a schematic diagram of a specific implementation manner of an attack dynamic detection and identification method of the heterogeneous cross-domain system. As shown in fig. 1, the attack dynamic detection and identification method for the heterogeneous cross-domain system in this embodiment includes the specific steps of:
the embodiment combines a heterogeneous cross-domain system domain unit U under network attack 1 Domain unit U 2 8230and domain unit U n The state characteristics and the environment characteristics of the heterogeneous cross-domain system are designed, and a heterogeneous cross-domain system multi-source information fusion mechanism is designed;
generally, the structure of a heterogeneous cross-domain system domain unit can be divided into an interaction control layer, an information fusion network layer and a multi-sensor cooperative sensing layer. The network attack can directly attack multi-sensor sensing layer nodes, network nodes and control layer nodes. Meanwhile, due to heterogeneous multi-source, random diversity and complexity of network attack, a heterogeneous cross-domain system multi-sensor collaborative sensing layer, an information fusion network layer and a control layer are extremely vulnerable to unknown malicious attack threats.
The attack dynamic detection and identification method of the heterogeneous cross-domain system is realized based on a generation countermeasure Network (GAN). Fig. 2 is a diagram of a generation countermeasure network structure. The generation countermeasure network is an unsupervised learning method for mutually playing games and learning through two neural networks, and is composed of a generation network G and a discrimination network D. The generation of the countermeasure network aims at training a generation network G, and multi-source fusion state data M (U) under real attack in a heterogeneous cross-domain system by utilizing a heterogeneous cross-domain system multi-source information fusion mechanism 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) In the random selection of variables as outputsIn, outputting data which is as close to a real sample as possible through training
At the same time, a discriminating network D is trained in order to generate data output by the network>
As its input, the discriminant network is trained to distinguish as much as possible the data generating the network output from the real samples. Iterative game and mutual antagonistic training are carried out on the two networks at the same time, parameters of the two networks are continuously adjusted, nash balance can be finally achieved, the purpose is to enable data output by the generated networks to be indistinguishable from real samples, and the generated data and the real data cannot be correctly distinguished by the judgment network.
The embodiment combines the characteristics of methods for generating the countermeasure network iterative game and generating the countermeasure, and utilizes the generation network G to perform dynamic fusion detection on the incompleteness, the asynchronism and the false performance of the state measurement information of the heterogeneous cross-domain system in the shortest time; and performing online fusion recognition on attack information by combining data under the training attack of the discrimination network D.
The heterogeneous cross-domain system under the network attack can be divided into domain units U 1 And domain unit U 2 8230the domain unit U n . Wherein, the domain unit U 1 Including status signature data U 1s (x 1s1 ,x 1s2 8230quickly) and environmental characteristic data U 1e (x 1e1 ,x 1e2 8230in which x 1s1 ，x 1s2 Presentation Domain Unit U 1 Characteristic of state of (1), x 1e1 ，x 1e2 Representation domain unit U 1 An environmental characteristic; domain unit U 2 Including status signature data U 2s (x 2s1 ,x 2s2 8230quickly) and environmental characteristic data U 2e (x 2e1 ,x 2e2 8230), where x 2s1 ，x 2s2 Presentation Domain Unit U 2 Status feature, x 2e1 ，x 2e2 Presentation Domain Unit U 2 An environmental characteristic; domain unit U n Including status feature data U ns (x ns1 ,x ns2 8230quickly) and environmental characteristic data U ne (x ne1 ,x ne2 ,…)，x ns1 ，x ns2 Representation domain unit U n Status feature, x ne1 ，x ne2 Representation domain unit U n And (6) state characteristics.
In the embodiment, the state characteristic data and the environment characteristic data of the heterogeneous cross-domain system domain unit under the network attack are analyzed, a heterogeneous cross-domain system multi-source information fusion mechanism is established, and real multi-source fusion state data M (U) under the attack are generated 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) (ii) a In the embodiment, a group of data is generated through training based on a generating network G and a heterogeneous cross-domain system multi-source information fusion mechanism
Wherein +>
Domain units U generated for training respectively 1 ，U 2 ，…,U n Is based on the status characteristic data of (4), is based on the status characteristic data of (5)>
Domain units U generated for training respectively 1 ，U 2 ，…,U n The environmental characteristic data of (1). Will generate data->
And real state data M (U) under attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And repeatedly and iteratively comparing to realize dynamic fusion detection of imperfection, asynchronism and false detection of the measurement information in the shortest time.
Furthermore, the real data M (U) under attack is utilized in this example 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And generating data
Training a discrimination network D, performing iterative game and mutual antagonistic training simultaneously through the generation network G and the discrimination network D, wherein the optimization target function is->
D and G respectively represent data sequences of the generation network and the judgment network, and come from the actual sensing data distribution of the system. Finally reach nash equilibrium, thus realize the online fusion recognition of the attack information, receive the true compound attack information A (a) 1 ,a 2 8230; e.g. a) 1 Represents a double lasso attack, a 2 Indicating a spoofing attack. />
Claims (1)
1. An attack dynamic detection and identification method of a heterogeneous cross-domain system comprises the heterogeneous cross-domain system under network attack, a multi-source information fusion mechanism of the heterogeneous cross-domain system, a generation network G module, a judgment network D module, a distributed fusion dynamic detection mechanism of the heterogeneous cross-domain system and an attack information on-line fusion identification mechanism; the method is characterized by comprising the following steps:
step 1: designing a heterogeneous cross-domain system multi-source information fusion mechanism by combining domain unit state characteristics and environment characteristics of the heterogeneous cross-domain system under network attack;
step 2: by combining the characteristics of methods for generating a countermeasure network iterative game and generating countermeasures, the G module of the generation network is utilized to carry out dynamic fusion detection on the incompleteness, asynchronism and false performance of the measurement information in the shortest time;
and step 3: on-line fusion recognition is carried out on attack information by judging state data under the training attack of a network D module;
the heterogeneous cross-domain system under the network attack is composed of a plurality of heterogeneous domain units and comprises different state characteristic information and environment characteristic information; the network attack comprises double lasso attack and decoy attack; specifically, the heterogeneous cross-domain system under the network attack can be divided into domain units U 1 Domain unit U 2 8230and domain unit U n (ii) a Wherein, the domain unit U 1 Including status signature data U 1s (x 1s1 ,x 1s2 8230quickly) and environmental characteristic data U 1e (x 1e1 ,x 1e2 8230), where x 1s1 ，x 1s2 Presentation Domain Unit U 1 Characteristic of state of (1), x 1e1 ，x 1e2 Presentation Domain Unit U 1 An environmental characteristic; domain unit U 2 Including status feature data U 2s (x 2s1 ,x 2s2 8230;) and environmental characteristic data U 2e (x 2e1 ,x 2e2 8230in which x 2s1 ，x 2s2 Representation domain unit U 2 Status feature, x 2e1 ，x 2e2 Representation domain unit U 2 An environmental characteristic; domain unit U n Including status feature data U ns (x ns1 ,x ns2 8230quickly) and environmental characteristic data U ne (x ne1 ,x ne2 ,…)，x ns1 ，x ns2 Presentation Domain Unit U n Status feature, x ne1 ，x ne2 Representation domain unit U n An environmental characteristic;
the heterogeneous cross-domain system multi-source information fusion mechanism is established by utilizing state characteristic data and environment characteristic data of a heterogeneous cross-domain system domain unit under network attack, and the generated real state data under the network attack is M (U) 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne )；
The generation network G module is based on a heterogeneous cross-domain system multi-source information fusion mechanism and is used for generating real state data M (U) under attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) Randomly selecting variables as input, and obtaining the variables through training;
the discrimination network D module is real state data M (U) under attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And generating data for network output
As input, obtained by training; wherein it is present>
Domain units U generated for training respectively 1 ，U 2 ，…,U n Status characteristic data of (a), based on the status of the device>
Domain units U generated for training respectively 1 ，U 2 ，…,U n The environmental characteristic data of (1);
the heterogeneous cross-domain system distributed fusion dynamic detection mechanism is combined with the characteristics of a method for generating a confrontation network iterative game and a confrontation, and a group of data is generated through training based on a generation network G module and a heterogeneous cross-domain system multi-source information fusion mechanism
The real state data M (U) under the attack is matched with the real state data 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) Repeated iterative comparison is carried out to realize dynamic fusion detection of imperfection, asynchronism and false detection of the measurement information in the shortest time;
the attack information online fusion recognition mechanism utilizes real state data M (U) under attack 1s ,U 1e ,U 2s ,U 2e ,…,U ns ,U ne ) And generating data
Training a discrimination network D, performing iterative game and mutual antagonistic training simultaneously through the generation network G and the discrimination network D, wherein the optimization target function is->
D, G respectively represents data sequences of a generation network and a judgment network, and comes from actual sensing data distribution of the system; finally reach nash equilibrium, thus realize the online fusion recognition of the attack information, receive the true compound attack information A (a) 1 ,a 2 8230), wherein the attack information includes a double lasso attack a 1 Trick attack a 2 。/>
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210408479.7A CN114726636B (en) | 2022-04-19 | 2022-04-19 | Attack dynamic detection and identification method for heterogeneous cross-domain system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210408479.7A CN114726636B (en) | 2022-04-19 | 2022-04-19 | Attack dynamic detection and identification method for heterogeneous cross-domain system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114726636A CN114726636A (en) | 2022-07-08 |
| CN114726636B true CN114726636B (en) | 2023-04-07 |
Family
ID=82243974
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210408479.7A Active CN114726636B (en) | 2022-04-19 | 2022-04-19 | Attack dynamic detection and identification method for heterogeneous cross-domain system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114726636B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158390A (en) * | 2021-04-29 | 2021-07-23 | 北京邮电大学 | Network attack traffic generation method for generating countermeasure network based on auxiliary classification |
| CN113505855A (en) * | 2021-07-30 | 2021-10-15 | 中国科学院计算技术研究所 | Training method for anti-attack model |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598400B (en) * | 2019-08-29 | 2021-03-05 | 浙江工业大学 | Defense method for high hidden poisoning attack based on generation countermeasure network and application |
| CN111163472B (en) * | 2019-12-30 | 2022-10-04 | 浙江工业大学 | Signal identification attack defense method based on generative countermeasure network |
| US11483318B2 (en) * | 2020-01-07 | 2022-10-25 | International Business Machines Corporation | Providing network security through autonomous simulated environments |
-
2022
- 2022-04-19 CN CN202210408479.7A patent/CN114726636B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158390A (en) * | 2021-04-29 | 2021-07-23 | 北京邮电大学 | Network attack traffic generation method for generating countermeasure network based on auxiliary classification |
| CN113505855A (en) * | 2021-07-30 | 2021-10-15 | 中国科学院计算技术研究所 | Training method for anti-attack model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114726636A (en) | 2022-07-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Deep learning based attack detection for cyber-physical system cybersecurity: A survey | |
| Liu et al. | An intrusion detection model with hierarchical attention mechanism | |
| Chen et al. | Collaborative spectrum sensing for illegal drone detection: A deep learning-based image classification perspective | |
| US20220094710A1 (en) | Detection of cyber attacks targeting avionics systems | |
| Kamhoua et al. | Game theory and machine learning for cyber security | |
| Kheddar et al. | Deep transfer learning for intrusion detection in industrial control networks: A comprehensive review | |
| Kheddar et al. | Deep transfer learning applications in intrusion detection systems: A comprehensive review | |
| Elsayed et al. | AdaptIDS: Adaptive intrusion detection for mission-critical aerospace vehicles | |
| Krishnan Sadhasivan et al. | A novel LWCSO-PKM-based feature optimization and classification of attack types in SCADA network | |
| Cengiz et al. | A novel intrusion detection system based on artificial neural network and genetic algorithm with a new dimensionality reduction technique for UAV communication | |
| Mohammed et al. | Comprehensive systematic review of intelligent approaches in UAV-based intrusion detection, blockchain, and network security | |
| CN114726636B (en) | Attack dynamic detection and identification method for heterogeneous cross-domain system | |
| CN112668697B (en) | Fuzzy test method and system for flight control parameters of unmanned aerial vehicle | |
| Han et al. | A unified anomaly detection methodology for lane-following of autonomous driving systems | |
| Jin et al. | Excitement surfeited turns to errors: Deep learning testing framework based on excitable neurons | |
| Gohari et al. | DEEP LEARNING-BASED INTRUSION DETECTION SYSTEMS: A COMPREHENSIVE SURVEY OF FOUR MAIN FIELDS OF CYBER SECURITY. | |
| Sun et al. | Application of Artificial Intelligence Technology in Honeypot Technology | |
| Su et al. | Intrusion Detection for Marine Meteorological Sensor Network | |
| Li et al. | An Anomaly Detection Approach Based on Integrated LSTM for IoT Big Data | |
| Yenduri et al. | Recent Advancements in Network and Cyber Security Using RNN | |
| Shao et al. | Anomaly recognition method of perception system for autonomous vehicles based on distance metric | |
| Ahsan et al. | A stacked ensemble learning ids model for software-defined vanet | |
| Anuradha et al. | Intrusion Detection System for Big Data Analytics in IoT Environment. | |
| Anitha et al. | Cybersecurity in Internet of Things Networks using Deep Learning Models | |
| Khoei et al. | A Comparative Assessment of Unsupervised Deep Learning Models for Detecting GPS Spoofing Attacks on Unmanned Aerial Systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |