International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 09 | Sept 2025
p-ISSN: 2395-0072
www.irjet.net
BMDNATP: Design of a Bioinspired Machine learning model for estimation of Dynamic Network Attacks via Traffic Pattern analysis Dr. Nikita Bahaley1 1Assistant Professor, Dept. of Information Technology, Pillai College of Arts, Commerce & Science (Autonomous),
New Mumbai, Maharashtra, India ---------------------------------------------------------------------***--------------------------------------------------------------------systems [5] involves continuous-time state measurement Abstract - The increasing number of cyber-attacks on
and monitoring of the triggering condition, both of which utilize a large amount of energy. The emergence of sampleddata ET systems has eased this need. In these procedures, merely periodic samples of the system are employed to evaluate the condition and monitor the ET state [6, 7]. Recent work has proposed a dynamic event triggering (DET) technique that integrates an extra dynamic variable in the ET condition. As illustrated in [1], the deployment of DETs resulted in fewer transmissions than so-called static ET approaches. Nonetheless, it is assumed in [8, 9, 10, 11, 12, 13] that measurements and events would be monitored in real time via Event-Triggered Dynamic Watermarking operations.
network systems has become a significant challenge for network security under real-time network scenarios. Deep learning models have proven to be effective in identifying network attacks. However, these models require a large amount of data for training, and their implementation can be computationally expensive when deployed on large-scale networks. To overcome these issues, this paper proposes a blockchain-based deep learning model that utilizes the advantages of blockchain to enhance the efficiency and security of network attack identification and mitigation. The proposed model uses a novel Proof-of-Wireless-Trust (PoWT) consensus algorithm to validate and secure the training data, and a customized Binary Cascaded Deep Learning Model (BCDLM) for training the model w.r.t. multiple attack signatures. The blockchain-based model is designed to detect and mitigate dynamic network attacks in real-time, thereby enhancing the security of network systems. The proposed model is evaluated using different network datasets.
The value of DET techniques is considerably boosted by integrating a mechanism for measuring and monitoring gathered data. A sampled-data DET framework for linear system stabilization has been proposed in [14, 15]. Riccatibased approaches [14, 15] complicate the controller design, and the control gains must be determined beforehand. In the context of ET control, there are frequently two techniques for developing unknown control and ET parameters: I emulation and (ii) co-design. In the emulation-based techniques, the control gain is first determined without taking the ET scheme into mind. In the second stage, the ET parameters are produced depending on the supplied control gain values. In co-design methodologies like Model Predictive Control (MPC) [16, 17, 18, 19, 20], the control gain and ET parameters are generated concurrently using a single framework. The ET parameter feasible zones are constrained by the initial selection of the control gains, which is a weakness of the emulation-based design via Neural Network-based Function Approximation Technique (NN FAT) [21, 22, 23, 24, 25]. The ET approach is consequently less effective in lowering the volume of transmissions. The application of the co-design technique for sampled-data dynamic event-triggered control deserves additional exploration. The functioning of NCSs may be dramatically reduced by assaults utilizing a number of malicious modification methods. If the adversary's impact is not adequately accounted for in the stability analysis, the NCS may become unstable in response to assaults. DDoS assaults are one of the most popular forms of cyberattacks (DoS) (DoS). In DoS, the opponent typically blocks transmission channels. Despite the fact that sensor/control packet losses
Key Words: Network, Attacks, Dynamic, Bioinspired, Machine, BFO
1.INTRODUCTION Stability is an essential component in the design and analysis of networking & control systems. In networked control systems (NCS), the restricted on-board energy resources provided to the system demand the implementation of solutions that decrease processing, such as status monitoring and control input update [1]. Due to the advent of cyberattacks, the resistance of NCSs (networked control systems) to different forms of assault has attracted substantial attention and is currently one of the most serious concerns in NCSs via Adaptive Dynamic Programming-based Optimal event-triggered NN (ADP ONN) controller [2]. For a long time, stability analysis and control design have relied on continuous-time state measurements and control approaches [3]. Researchers were required to utilize sampled-data techniques [4] in NCSs owing of physical limits such as energy and network capacity, which needed frequent state assessment and control adjustments. In recent years, the use of event-triggered (ET) techniques to further decrease processing has gained attention. In the ET systems, data updates and transfers are only sent out if specific criteria are met. The detection of events in various ET
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 437