From Challenges to Opportunities: A Comprehensive Study of AI-based In-Vehicle Intrusion Detection Systems
Year: 2024
Type: conference-paper
Abstract: While significant research has been conducted onML-based In-Vehicle Intrusion Detection Systems (IV-IDS), thepractical application of these systems needs further refinement.The safety-critical nature of IV-IDS calls for precise and objectiveevaluation and feasibility assessment metrics. This paperresponds to this need by conducting a rigorous ML-based IVIDSanalysis. We offer a thorough review of recent automotiveforensics studies spotlighting the constraints relevant to Invehiclesnetworks and the associated security/safety requirementsto reveal the current gaps in the existing literature. By addressingthe limitations of AI in IV-IDS, this paper contributes to theexisting research corpus and defines pertinent baseline metricsfor in-vehicle networked systems. Essentially, we reconcile therequirements of real-world autonomous vehicles with those ofthe security domain, enabling an assessment of the viability ofAI-based intrusion detection systems.
Source: HAL
Institutions: Institut de Recherche Technologique SystemX, Valeo (France)
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4392091835
