Detection of Web based-Attacks Using Artificial Intelligence

Abstract

The cyber-physical systems (cps) have improved significantly in many dynamic applications. Yet, these systems are seriously threatened by cyberattacks. Cyber-attacks, in contrast to flaws that result from errors in cyber-physical systems, are premeditated and covert. Some of these assaults, also known as deception attacks, manipulate certain cyber components or introduce phoney data from sensors or controllers into the system to damage data or introduce misleading information. Inability to recognise these attacks if the system is not aware of them could cause performance issues or even total system failure. To identify these attacks in these systems, algorithms must be altered.It should be mentioned that since the data generated by these systems is produced in such vast quantities, with such a broad diversity, and at such a quick rate, the application of machine learning algorithms is needed to make it easier to analyse, appraise, and uncover hidden patterns in the data. The CPS is modelled in this study as a network of moving agents, one of which acts as the leader and the rest of which are submissive to it. The approach proposed in this study makes advantage of the deep neural network's structural features for the detection stage, which ought to alert the system to the attack's presence in its first stages. In a leader-follower mechanism, to isolate the problematic agent, The application of resilient control methods in a network has been researched. The control system in the previously discussed control approach use a deep neural network to identify an attack before isolating the attacking agent and then isolating the misbehaving agent using a reputation mechanism. Experimental investigations have shown that deep learning algorithms may identify assaults more precisely than traditional techniques, as well as simplify, be proactive, cost-effective, and greatly enhance cyber security.