• Tidak ada hasil yang ditemukan

MACHINE LEARNING APPROACH FOR DETECTION, ESTIMATION AND COMPENSATION OF MALICIOUS ATTACKS IN NON LINEAR CYBER CRIME PHYSICAL SYSTEMS

N/A
N/A
Protected

Academic year: 2025

Membagikan "MACHINE LEARNING APPROACH FOR DETECTION, ESTIMATION AND COMPENSATION OF MALICIOUS ATTACKS IN NON LINEAR CYBER CRIME PHYSICAL SYSTEMS"

Copied!
5
0
0

Teks penuh

(1)

Abstract— This paper proposes an adaptive intelligent and classical approach for monitoring cyber-attacks on CPS inputs and smart manufacturing systems controlled by communications networks. A class of nonlinear n-order systems is referred to in this study as a CPS system, but only in the forwarding channel in the presence of cyber-attacks. An intelligent control system for cyber- attacks has been developed. A supervised learning optimization algorithm an SVM (Support Vector Machine) has been built in a standard, non-linear system that relies on changing the amount control mechanism to compensate for attack effects and control system performance and track applications. The proposed technique also uses security encryption and decryption. The results of the simulation indicate the feasibility and efficacy of the methodology suggested.

Index Terms— cyber security, simulation, attacks, SVM

I. INTRODUCTION

CYBER Physical Systems are system integration, networking and physical processes (CPS). A CPS consists of machine objects that collaborate with and processes of the real world.

Cyber Physical Systems have a similar internet of things relative (IoT), but they are further apart because of the special association characteristic with physical entities [1-3]. Several CPS examples occur in the case of autonomous vehicles, robotic operation, intelligent cities, intelligent power grid, intelligent manufacturing and embedded medical devices. The CPS's omnipresence and omnipresence are an important characteristic of the omnipresent development technology. Manufacturing systems are currently supported in the form of industrial IoT [4–

7] and SOA (service-oriented architectures). Some CPSs rely on ad hoc networks or the internet for message exchange and signal monitoring. This vulnerability makes the system vulnerable to network domain attacks. CPS attacks can cause severe damage.

This attack does not inherently begin in the virtual world and will occur in the physical environment. CPS is vulnerable to attacks of all components. Knowledge and cyber security techniques alone are also not sufficient to ensure that CPS performs properly [8]. Systems could be used to add security shields to CPS information. Such structures can provide robustness of attack. They may also be used in a wider breach mitigation and compensation scheme. Even cyber attacks on CPS can lead to defects and failures of physical devices. One research problem in these systems is the automatic compensation of (deliberate) failure effects and maintenance of system performance at a proper stage. Normally, target system processes or sensors for attacks or failures. An attack detection and control framework aims to improve system vailability through the design of control algorithms that preserve stability

and efficiency in case of defects (CPS attacks). This article proposes a classic smart control approach for non-linear scalar

CPS attacks. Cyber-attacks are predicted in the future in nonlinear CPS. The built-in control system contains a nonlinear controller based on the variable structure (VS) method and a Smart Attack Effect Estimator (GRBFNN). The VS control procedure is a robust control technique involving two main steps, which involve the choice of the appropriate switching surface and a robust control input. The GRBFNN estimator provides on-line prediction of possible attacks and the adaptive neural VS controller offsets the physical device effect of such attacks and monitors performance for monitoring and tracking purposes. The Smart Estimator Adaptation Act comes from a Lyapunov stability analysis. This theorem thus ensures an asymptotically stable structure.

II.REFERENCE WORKS

In the past ten years, CPS research has made considerable efforts from a variety of angles such as applications, safety, vulnerability, etc. Electricity, transport, health and manufacturing have become more and more part of the use of CPS. CPS has a broad description of SCADA systems for vital infrastructure such as intelligent [9–11] and smart grid [12].

CPSs also include conventional power grids, water grids, ICSs and intelligent vehicles. In order to boost fuel economy, safety and comfort, a system of small control systems is incorporated in modern smart cars [13, 14]. General approaches to analyzing the vulnerability of CPS can be used for evaluating the effects of such threats such as deceptions [11], denial of service [15], attacks [15], responses [17], covert assaults [18], false data injections [19, 20], etc.

The state evaluation of frequency components to whether they are vulnerable to faults, defects or assaults has been an important concern in recent years. Onlooker techniques are typically used to design error detection systems. The discrepancy between system output and output is referred to as residue [21]. Residual methods for fault diagnosis using a linear observation deck were widely used. Its exponential efficiency and responsiveness to (naturally limited) disturbances of modeling is severely restricted between linear observatories [22]. Rigorous sliding mode observations are designed to avoid and estimate online CPSs within a limited period of time [23] for state and sensor attacks. The Attack Evaluation and Compensation was also undertaken for restricted modeling errors affecting an electricity network power system of the United States Western Electricity Coordinating Council (WECC) [24]. An appropriate governance system is needed to detect and then run CPSs safely against cyber-attacks. Attack assessment and monitoring of results are important to ensure robustness and stability in

MACHINE LEARNING APPROACH FOR DETECTION, ESTIMATION AND COMPENSATION OF MALICIOUS ATTACKS IN NON LINEAR CYBER CRIME

PHYSICAL SYSTEMS

A.Praveena

1

, ShreeHarsha.M

2

, Rohith.R

3

, Niaz Ahmed.M

4

1

Assistant Professor, Department of CSE, Jansons Institute of Technology, Coimbatore

2,3,4

Final year B.E.(CSE), Jansons Institute of Technology, Coimbatore

(2)

IJRAR21B1430 International Journal of Research and Analytical Reviews (IJRAR) www.ijrar.org 742 these scenarios. Many studies have already been carried out on

fault identification, perception and tolerance management [21].

However, less work is being done on linear cps, for example for power grids [25, 26], water grids [27–29], etc. from a security point of view. The network of sensors and the security of computers are used as prevention mechanisms and the focus is not on how the control system can proceed when CPS is attacked [30]. There are two defence layers for safe CPSF operation. First, to prevent attacks on CPS, software security layers are introduced. It can achieve safety goals, honesty, availability and trust. The second level is to counter the effect of attacks when the adversary/attacker reaches the CPS. The mechanism control/compensator can be used. Research has been carried out on cyber-attack identification [31-33], isolation and survival problems for CPS [8, 17, 22]. We know that there is no study on the restoration of online attacks. In [34] the paper focused on a second level of cyber-attack security, particularly when the cyber-attack input was rebuilt. The WECC network power system in attack [23] is built as a linear system subject to unknown inputs that modify the network attack and sensor attack. Sliding mode observers are programmed to monitor and replay all attacks in the shortest time possible. The extension of [23] takes place in [24]. The attack is compensated if the performance tracks a given direction. In [35], an intrusion detection and compensation system is established based on the identification of the computer for fighting covert attacks. Errors are obtained in the performance assessment during the system activity learning stage, and system behavior is monitored to verify if it deviates significantly from the expected outputs. A compensating controller can, after the attack is detected, also intervene and replace the traditional controller. [36] The control strategy is provided in rotary portal inputs and outputs for the accommodating control and compensation of cyber-attacks. The malicious attacks on both signals, input control and the output sensor, are thought to be a Denial of Service (DoS).

This paper explores many classic and intelligent approaches in terms of robustness and efficacy for controlling cyber-attacks. It is also noteworthy that the majority of studies are carried out on linear dynamic systems. The linearization approach is then used to investigate nonlinear dynamic systems based on the extension of the Taylor series. Many studies focus on linear or linear processes. Often this simplification is not correct. We focus on non-linear systems and in this article there are no sequential assumptions. Moreover, online evaluation and compensation of attacks plus external disturbances have been studied.

III. EXISTING SYSTEM

In this analysis, the control system aims to improve the security and reliability of CPS based on the assumption that it is susceptible to attack. This is achieved through the development of control algorithms to ensure stability and productivity in the face of attacks. To fight covert attacks, a system recognition-based intrusion prevention system and compensation tool is proposed. Output prediction errors are collected during the system's learning process, and system behaviour is tracked to see if it deviates dramatically from the expected output. The accounted controller is designed to interact with and disable the conventional controller when the attack has been detected.

IV. PROPOSED SYSTEM

The purpose of the control system in this study is to increase the safety and reliability of CPS, provided that it is vulnerable to attacks. This is done through the design of control algorithms which can maintain stability and efficiency in attack scenarios.

Fig 1. Modules of the proposed system Hardware Requirements

 Hardware - Pentium III I3

 Processor Speed - 2.1 GHz

 RAM - 4GB

 Hard Disk - 160 GB

Software Requirements

 Operating System - Windows 7/8/10

 Technology -PYTHON

MODULES

DATA COLLECTION

The process of gathering, measuring and evaluating accurate insights on research using standard validated techniques is known as the data collection. On the basis of collected data, a researcher will assess his hypothesis. In most cases data collection is, irrespective of the field of study, the primary and most critical phase in research. The data is collected and the data is clustered fluently Fuzzy logical concepts can be used to group multidimensional data, assigning membership from 0 to 100 percent per point in each cluster core. This can be very effective compared to conventional hard threshold clusters with a crisp, exact mark allocated to each object.

(3)

Fig 2. Data Collection AES AND DES

The world is literally struggling with what tends to be a cyber- attack – a complex, multi-vectored attack with clear cyber pandemic characteristics. A cyber assault can be used to maliciously disable computers, steal data, or use a compromised computer as a launch pad for other attacks. To conduct a cyber- attack, cybercriminals use a range of tools, including malware, phishing, ransom ware, and denial of service, among others.

These cyber-attacks can be rectified using AES AND DES for security purposes. The AES encryption algorithm describes various transformations to be carried out on data in an array.

The first step is to add the data to the array — after which, over several encryption ranges, cypher transformations are repeated.

The first transformation in the AES encryption chip is data replacement with a replacement table; the second transformation is data row replacement; and the third one is column mixing.

The last transformation is done with a different part of the encryption key on each column. Longer keys need to complete more rounds. In this AES henon map for used for encryption.

The aim of DES (Data Encryption Standard) algorithm is to provide a standard method for protecting commercially sensitive and unclassified data. In the same key for encryption and decryption.

Fig 3. Encryption using Henon map

ATTACK SIMULATION

The simulation of attacks shows you how your network and security measures can function against real-world attack scenarios. The method proposed makes sophisticated simulation violations and attacks to authenticate and control safety positions. Measure safety threats to identify vulnerable areas and track safety improvements. Recommendations for remedying security gaps and optimizing safety measures.

Security specialists with the ability to leverage threat intelligence through life cycles of attacks and recognize occult threat vectors. Proven expertise in developing governance

hunting simulations based on the result. Advanced ability to simulate threats to assess and motivate security programmers to minimize hazards.

Fig 4. Attack detected

Fig 5. No Attack

IMPLEMENTATION OF AI TO DETECT THE ATTACK

Computational Intelligence is a form of artificial intelligence (CI). Other nature-inspired techniques used in various fields which provide versatile decision-making frameworks for complex environments including cyber-security applications.

Artificial intelligence tools can be used to filter out noise and unnecessary data, as well as to help security experts understand the cyber environment and identify unusual behavior. By integrating with AI the accuracy level of attack can be determined.

Fig 6. Accuracy level for Attack detected

(4)

IJRAR21B1430 International Journal of Research and Analytical Reviews (IJRAR) www.ijrar.org 744 Fig 7. Accuracy level for Attack not detected

V. CONCLUSIONS AND FUTURE WORK The proposed concept works on removing cyber-attacks by performing encryption and decryption algorithm where output efficiency is increased by using AES and DES. Here like firewall etc, we have strengthened the security by using efficient encryption AES techniques. The accuracy of attack detected and not detected are printed by integrating with AI.

REFERENCES

[1] L. Monostori, B. Kad´ ar, T. Bauernhansl, S. Kondoh,

´ S. Kumara, G. Reinhart, O. Sauer, G. Schuh, W. Sihn, and K. Ueda, “Cyber-physical systems in manufacturing,”

CIRP Annals, vol. 65, no. 2, pp. 621–641, 2016.

[2] A. Humayed, J. Lin, F. Li, and B. Luo, “Cyber- physical systems securitya survey,” IEEE Internet of Things Journal, vol. 4, no. 6, pp. 1802–1831, 2017.

[3] Y. Ashibani and Q. H. Mahmoud, “Cyber physical systems security: Analysis, challenges and solutions,”

Computers & Security, vol. 68, pp. 81–97, 2017.

[4] D. Ding, Q.-L. Han, Z. Wang, and X. Ge, “A survey on model-based distributed control and filtering for industrial cyber-physical systems,” IEEE Transactions on Industrial Informatics, vol. 15, no. 5, pp. 2483–2499, 2019.

[5] A. Bonci, M. Pirani, and S. Longhi, “Tiny cyber- physical systems for performance improvement in the factory of the future,” IEEE Transactions on Industrial Informatics, vol. 15, no. 3, pp. 1598–1608, 2018.

[6] Y. Zhang, Z. Guo, J. Lv, and Y. Liu, “A framework for smart production-logistics systems based on cps and industrial iot,” IEEE Transactions on Industrial Informatics, vol. 14, no. 9, pp. 4019–4032, 2018.

[7] A. Arabsorkhi, M. S. Haghighi, and R. Ghorbanloo,

“A conceptual trust model for the internet of things interactions,” in 8th International Symposium on Telecommunications (IST), 2016, pp. 89–93.

[8] F. Pasqualetti, F. Dorfler, and F. Bullo, “Control- theoretic methods for cyberphysical security: Geometric principles for optimal cross-layer resilient control systems,” IEEE Control Systems, vol. 35, no. 1, pp. 110–

127, 2015.

[9] A. Jolfaei and K. Kant, “Privacy and security of connected vehicles in intelligent transportation system,”

in 2019 49th Annual IEEE/IFIP International Conference on Dependable Systems and Networks–Supplemental Volume (DSN-S). IEEE, 2019, pp. 9–10.

[10] A. Jolfaei, K. Kant, and H. Shafei, “Secure data streaming to untrusted road side units in intelligent transportation system,” in 2019 18th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/13th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDataSE). IEEE, 2019, pp.

793–798.

[11] S. G. Ezabadi, A. Jolfaei, L. Kulik, and R. Kotagiri,

“Differentially private streaming to untrusted edge servers in intelligent transportation system,” in 2019 18th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/13th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDataSE). IEEE, 2019, pp.

781–786.

[12] A. Mohammadali, M. S. Haghighi, M. H. Tadayon, and A. Mohammadi-Nodooshan, “A novel identity-based key establishment method for advanced metering infrastructure in smart grid,” IEEE Transactions on Smart Grid, vol. 9, no. 4, pp. 2834–2842, 2016.

[13] A. Humayed and B. Luo, “Cyber-physical security for smart cars: taxonomy of vulnerabilities, threats, and attacks,” in Proceedings of the ACM/IEEE Sixth Inter.

Conf. on Cyber-Physical Systems. ACM, 2015.

[14] H.-K. Kong, M. K. Hong, and T.-S. Kim, “Security risk assessment framework for smart car using the attack tree analysis,” Journal of Ambient Intelligence and Humanized Computing, vol. 9, no. 3, pp. 531–551, 2018.

[15] S. Amin, A. A. Cardenas, and S. S. Sastry, “Safe and ´ secure networked control systems under denial-of- service attacks,” in International Workshop on Hybrid Systems: Computation and Control. Springer, 2009, pp.

31–45.

[16] J.-Y. Keller and D. Sauter, “Monitoring of stealthy attack in networked control systems,” in Control and FaultTolerant Systems, 2013 Conference on. IEEE, 2013.

[17] Y. Mo and B. Sinopoli, “Secure control against replay attacks,” in Communication, Control, and Computing, Annual Allerton Conf. on. IEEE, 2009.

[18] A. O. de Sa, L. F. R. da Costa Carmo, and R. C. ´ Machado, “Covert attacks in cyber-physical control systems,” IEEE Transactions on Industrial Informatics, vol. 13, no. 4, pp. 1641–1651, 2017.

[19] Y. Mo, E. Garone, A. Casavola, and B. Sinopoli,

“False data injection attacks against state estimation in wireless sensor networks,” in Decision and Control (CDC), 2010 49th IEEE Conference on. IEEE, 2010, pp.

5967–5972.

(5)

[20] O. A. Beg, T. T. Johnson, and A. Davoudi,

“Detection of false-data injection attacks in cyber-physical

dc microgrids,” IEEE Transactions on industrial

informatics, vol. 13, no. 5, pp. 2693–2703, 2017.

Referensi

Dokumen terkait