A Smart Anomaly Detection Method in Cyber Physical Systems Using Machine Learning

A Smart Anomaly Detection Method in Cyber Physical Systems Using Machine Learning

P. Ramkumar, B. Shadaksharappa, R. Uma, J. Anitha Ruth, R. Valarmathi
DOI: 10.4018/979-8-3693-4159-9.ch021
OnDemand:
(Individual Chapters)
Available
$33.75
List Price: $37.50
10% Discount:-$3.75
TOTAL SAVINGS: $3.75

Abstract

Cyber-Physical systems (CPS) are recognized by a broad range of complicated multi-tasking fixings with good interaction which results in combining cyber areas within the actual physical world. Thinking about the substantial development of cyber physical methods as well as a result of the prevalent utilization of sensible communication and features equipment, brand new issues have emerged. With this regard, a brand new model of CPSs for an intelligent power grid are confronting various vulnerabilities and lots of attacks and threats. Anomaly detection is a crucial information evaluation undertaking as among the techniques for CPSs protection. As various anomaly detection techniques are provided, it's tough to evaluate the pros and cons of the strategies. In this particular chapter machine learning (ML) techniques for detection of anomalies are provided by way of a situation learning that shows the usefulness of machine learning methods at classifying false data injection (FDI) strikes.
Chapter Preview
Top

Introduction

In CPS, the devices, computational resources, and physical components are integrated through correspondence networks for remote control and monitoring, and the efficacy of the mixed approaches is demonstrated (Cicconi.P,2017) Some of the many benefits that have accrued from the use of these methods—self-healing, customer engagement, control, adaptive protection, etc.—include the need for practical tools for electric flow and improved functionality. Many difficulties arise, including issues with security, stability, reliability, predictability, and maintainability (Darwish, A, 2018) Cyber physical methods have many moving components, which makes them vulnerable from both a cyber and physical perspective; one of the most important challenges is overcoming this integration. Directed at either physical components or the cyber infrastructure, malicious attacks have disrupted method procedures or stolen esoteric details. Safe preparation, analysis, and storage of produced data is within the capabilities of machine learning (Kocabay, A.et.al, 2023). In this chapter, ML is used to talk about how to get predictions from product data after studying it.

Figure 1.

ML based CPS security

979-8-3693-4159-9.ch021.f01
Top

Proposed System

Cyber Physical System (CPS)

Different approaches could be identified based on CPS use; these approaches could include physically linked computation, interaction, social networking, and skilled resources; and physically linked physical tasks that rely on IT methods used to monitor and constrain the physical society. Look at Figure2 for a different take on CPSs (Jiang, J.R, 2018). Various descriptions of CPS are available, each focusing on a different aspect of the approaches. We will quickly go over them: cyber ability, integration, networking, dependability, automation, complexity, and reconfiguring. With cyber-physical approaches, physical components and cyber capabilities, such as sentinel networks, are integrated. The need for CPS dependability and security necessitates adaptive skills paired with advanced comment management systems.

Figure 2.

View of CPS

979-8-3693-4159-9.ch021.f02
Figure 3.

Cyber attacks in CPS

979-8-3693-4159-9.ch021.f03

We focus on security issues for the most part, but there are a number of challenges that cyber physical systems face from the many angles shown in Figure 3. This is because these systems rely on sentient interaction and intelligent sensors and tools. In the intelligent power grid, security is a major concern due to the increased reliance on cyber data by devices, which introduces new security holes. Excellent examples of CPSs, these devices provide the framework needed to deal with new challenges, and they have intensive correspondence capacities (Ding, D.et.al, 2019). A number of technological developments and a skyrocketing demand for electricity have prompted the creation of intelligent power grid. It is clear from this that a comprehensive plan is necessary to address the issue at hand, since we need to understand the effects of the hits and evaluate the efficacy of potential responses.

Complete Chapter List

Search this Book:
Reset